From 0b39f8c616f8cfb149572a3f15e470b313415271 Mon Sep 17 00:00:00 2001 From: Subramanyam Challa <46422196+SubramanyamChalla24@users.noreply.github.com> Date: Fri, 7 Jun 2024 01:15:42 -0600 Subject: [PATCH] Updated version (#264) * added data extraction files * additional script files * parsing and processing file run_first.py * main.py * removed cohere dependency , update codes . (#261) (cherry picked from commit a1cc1706ffab7d2527a49d4061b20ae792ba9cc3) * removed tkinter , removed cohere dependent code * added resumes , jd with new name format --- ...f9e508eff-a083-4e2d-8b1e-577cbc8f50fc.json | 265 -------------- ...fcccf72e0-5f20-4aa8-8679-d91b720b7247.json | 261 ------------- ...f1fb6435c-531a-4f04-84c2-c4e73e1f1a3f.json | 283 -------------- ...f0b613898-1183-49c5-ad07-e03bd9af72e5.json | 247 ------------- ...fb36fac00-1d60-49ba-a9e4-33477c928e98.json | 324 ---------------- ...fbef9d9b1-fa14-4aa0-8111-6033cceacb5a.json | 260 ------------- ...f04c228e3-04a4-4970-a149-a3cffc92aee8.json | 345 ------------------ ...f5dde9183-5af4-4107-91e1-7b2d66fec490.json | 304 --------------- ...f817fb796-14b9-4819-a224-7eed66b3ce04.json | 289 --------------- resume_matcher/__init__.py | 0 resume_matcher/dataextractor/DataExtractor.py | 221 +++++++++++ .../dataextractor/KeyTermExtractor.py | 154 ++++++++ resume_matcher/dataextractor/TextCleaner.py | 124 +++++++ resume_matcher/dataextractor/__init__.py | 0 resume_matcher/main.py | 39 ++ resume_matcher/run_first.py | 59 +++ resume_matcher/scripts/__init__.py | 0 resume_matcher/scripts/get_score.py | 77 ++++ resume_matcher/scripts/logger.py | 148 ++++++++ resume_matcher/scripts/parser.py | 55 +++ resume_matcher/scripts/processor.py | 54 +++ resume_matcher/scripts/utils.py | 147 ++++++++ scripts/similarity/get_similarity_score.py | 1 - 23 files changed, 1078 insertions(+), 2579 deletions(-) delete mode 100644 Data/Processed/JobDescription/JobDescription-job_desc_front_end_engineer.pdf9e508eff-a083-4e2d-8b1e-577cbc8f50fc.json delete mode 100644 Data/Processed/JobDescription/JobDescription-job_desc_full_stack_engineer.pdfcccf72e0-5f20-4aa8-8679-d91b720b7247.json delete mode 100644 Data/Processed/JobDescription/JobDescription-job_desc_java_developer.pdf1fb6435c-531a-4f04-84c2-c4e73e1f1a3f.json delete mode 100644 Data/Processed/JobDescription/JobDescription-job_desc_product_manager.pdf0b613898-1183-49c5-ad07-e03bd9af72e5.json delete mode 100644 Data/Processed/Resumes/Resume-alfred_pennyworth_pm.pdfb36fac00-1d60-49ba-a9e4-33477c928e98.json delete mode 100644 Data/Processed/Resumes/Resume-barry_allen_fe.pdfbef9d9b1-fa14-4aa0-8111-6033cceacb5a.json delete mode 100644 Data/Processed/Resumes/Resume-bruce_wayne_fullstack.pdf04c228e3-04a4-4970-a149-a3cffc92aee8.json delete mode 100644 Data/Processed/Resumes/Resume-harvey_dent_mle.pdf5dde9183-5af4-4107-91e1-7b2d66fec490.json delete mode 100644 Data/Processed/Resumes/Resume-john_doe.pdf817fb796-14b9-4819-a224-7eed66b3ce04.json create mode 100644 resume_matcher/__init__.py create mode 100644 resume_matcher/dataextractor/DataExtractor.py create mode 100644 resume_matcher/dataextractor/KeyTermExtractor.py create mode 100644 resume_matcher/dataextractor/TextCleaner.py create mode 100644 resume_matcher/dataextractor/__init__.py create mode 100644 resume_matcher/main.py create mode 100644 resume_matcher/run_first.py create mode 100644 resume_matcher/scripts/__init__.py create mode 100644 resume_matcher/scripts/get_score.py create mode 100644 resume_matcher/scripts/logger.py create mode 100644 resume_matcher/scripts/parser.py create mode 100644 resume_matcher/scripts/processor.py create mode 100644 resume_matcher/scripts/utils.py diff --git a/Data/Processed/JobDescription/JobDescription-job_desc_front_end_engineer.pdf9e508eff-a083-4e2d-8b1e-577cbc8f50fc.json b/Data/Processed/JobDescription/JobDescription-job_desc_front_end_engineer.pdf9e508eff-a083-4e2d-8b1e-577cbc8f50fc.json deleted file mode 100644 index de1f8821..00000000 --- a/Data/Processed/JobDescription/JobDescription-job_desc_front_end_engineer.pdf9e508eff-a083-4e2d-8b1e-577cbc8f50fc.json +++ /dev/null @@ -1,265 +0,0 @@ -{ - "bi_grams": "[Job Description, End Engineer, Tech Solutions, Solutions San, San Francisco, Tech Solutions, build products, solve complex, complex problems, improve people, End Engineer, dynamic team, San Francisco, Job Description, End Engineer, developing scalable, userfriendly web, web applications, successful candidate, modern JavaScript, JavaScript frameworks, libraries HTML, HTML CSS, responsive design, design principles, contribute significantly, user interfaces, web applications, \u2022Develop new, new userfacing, userfacing features, modern JavaScript, JavaScript frameworks, frameworks like, like Reactjs, Reactjs Vuejs, \u2022Build reusable, reusable code, future use, technical feasibility, UX designs, \u2022Optimize application, maximum speed, user input, backend services, team members, \u20222 years, End Developer, similar role, web markup, markup including, including HTML5, HTML5 CSS3, modern JavaScript, JavaScript programming, libraries like, like jQuery, modern frontend, frontend build, build pipelines, popular frontend, frontend frameworks, React Vue, code versioning, versioning tools, Computer Science, Science Engineering, related field, \u2022Competitive salary, salary package, \u2022Health dental, vision insurance, \u2022Retirement savings, savings plan, \u2022Professional development, development opportunities, \u2022Flexible work, work hours, Tech Solutions, equal opportunity, opportunity employer, celebrate diversity, inclusive environment, brief explanation, relevant experience]", - "clean_data": "Job Description Front End Engineer 2 Years of Experience\nTech Solutions San Francisco CA USA\nAbout Us\nAt Tech Solutions we are on a mission to build products that solve complex problems and improve people\u2019s\nlives We are seeking a talented Front End Engineer to join our dynamic team in San Francisco\nJob Description\nWe are looking for a Front End Engineer with at least 2 years of experience in developing scalable and\nuserfriendly web applications The successful candidate will be proficient in modern JavaScript frameworks\nand libraries HTML CSS and responsive design principles This role will contribute significantly to the\ncreation and implementation of user interfaces for our web applications\nResponsibilities\n\u2022Develop new userfacing features using modern JavaScript frameworks like Reactjs Vuejs or Angu\nlarjs\n\u2022Build reusable code and libraries for future use\n\u2022Ensure the technical feasibility of UI/UX designs\n\u2022Optimize application for maximum speed and scalability\n\u2022Assure that all user input is validated before submitting to backend services\n\u2022Collaborate with other team members and stakeholders\nRequirements\n\u20222 years of experience as a Front End Developer or similar role\n\u2022Proficiency in web markup including HTML5 CSS3\n\u2022Knowledge of modern JavaScript programming and experience with libraries like jQuery\n\u2022Familiarity with modern frontend build pipelines and tools\n\u2022Experience with popular frontend frameworks such as React Vue or Angular\n\u2022Familiarity with code versioning tools such as Git\n\u2022Degree in Computer Science Engineering or a related field\n1 Benefits\n\u2022Competitive salary package\n\u2022Health dental and vision insurance\n\u2022Retirement savings plan\n\u2022Professional development opportunities\n\u2022Flexible work hours\nTech Solutions is proud to be an equal opportunity employer We celebrate diversity and are committed\nto creating an inclusive environment for all employees\nHow to Apply\nTo apply please submit your resume and a brief explanation of your relevant experience to \n2", - "emails": [ - "hiring@techsolutions.com" - ], - "entities": [ - "Git", - "JavaScript", - "HTML CSS", - "UI/UX", - "Tech Solutions", - "Front End Developer", - "Computer Science Engineering", - "San Francisco", - "Front End Engineer", - "Us", - "Job Description Front End Engineer" - ], - "experience": "Experience \n Tech Solutions San Francisco CA USA \n About Us \n At Tech Solutions we are on a mission to build products that solve complex problems and improve people \u2019s \n lives We are seeking a talented Front End Engineer to join our dynamic team in San Francisco \n Job Description \n We are looking for a Front End Engineer with at least 2 years of experience in developing scalable and \n userfriendly web applications The successful candidate will be proficient in modern JavaScript frameworks \n and libraries HTML CSS and responsive design principles This role will contribute significantly to the \n creation and implementation of user interfaces for our web applications \n Responsibilities \n \u2022Develop new userfacing features using modern JavaScript frameworks like Reactjs Vuejs or Angu \n larjs \n \u2022Build reusable code and libraries for future use \n \u2022Ensure the technical feasibility of UI / UX designs \n \u2022Optimize application for maximum speed and scalability \n \u2022Assure that all user input is validated before submitting to backend services \n \u2022Collaborate with other team members and stakeholders \n Requirements \n \u20222 years of experience as a Front End Developer or similar role \n \u2022Proficiency in web markup including HTML5 CSS3 \n \u2022Knowledge of modern JavaScript programming and experience with libraries like jQuery \n \u2022Familiarity with modern frontend build pipelines and tools \n \u2022Experience with popular frontend frameworks such as React Vue or Angular \n \u2022Familiarity with code versioning tools such as Git \n \u2022Degree in Computer Science Engineering or a related field \n 1 Benefits \n \u2022Competitive salary package \n \u2022Health dental and vision insurance \n \u2022Retirement savings plan \n \u2022Professional development opportunities \n \u2022Flexible work hours \n Tech Solutions is proud to be an equal opportunity employer We celebrate diversity and are committed \n to creating an inclusive environment for all employees \n How to Apply \n To apply please submit your resume and a brief explanation of your relevant experience to \n 2", - "extracted_keywords": [ - "Job", - "Description", - "Front", - "End", - "Engineer", - "Years", - "Experience", - "Tech", - "Solutions", - "San", - "Francisco", - "CA", - "USA", - "Tech", - "Solutions", - "mission", - "products", - "problems", - "people", - "End", - "Engineer", - "team", - "San", - "Francisco", - "Job", - "Description", - "End", - "Engineer", - "years", - "experience", - "web", - "applications", - "candidate", - "JavaScript", - "frameworks", - "HTML", - "CSS", - "design", - "principles", - "role", - "creation", - "implementation", - "user", - "interfaces", - "web", - "applications", - "Responsibilities", - "\u2022Develop", - "userfacing", - "features", - "JavaScript", - "frameworks", - "Reactjs", - "Vuejs", - "Angu", - "larjs", - "code", - "libraries", - "use", - "\u2022Ensure", - "feasibility", - "UI", - "UX", - "\u2022Optimize", - "application", - "speed", - "scalability", - "user", - "input", - "services", - "\u2022Collaborate", - "team", - "members", - "stakeholders", - "Requirements", - "\u20222", - "years", - "experience", - "End", - "Developer", - "role", - "\u2022Proficiency", - "web", - "markup", - "HTML5", - "CSS3", - "\u2022Knowledge", - "JavaScript", - "programming", - "experience", - "libraries", - "jQuery", - "\u2022Familiarity", - "frontend", - "pipelines", - "tools", - "\u2022Experience", - "frontend", - "frameworks", - "React", - "Vue", - "Angular", - "\u2022Familiarity", - "code", - "versioning", - "tools", - "Git", - "\u2022Degree", - "Computer", - "Science", - "Engineering", - "field", - "Benefits", - "salary", - "package", - "\u2022Health", - "vision", - "insurance", - "savings", - "plan", - "development", - "opportunities", - "work", - "hours", - "Tech", - "Solutions", - "opportunity", - "employer", - "diversity", - "environment", - "employees", - "resume", - "explanation", - "experience" - ], - "keyterms": [ - [ - "modern JavaScript framework", - 0.26131022145299865 - ], - [ - "End Engineer", - 0.13359819259474623 - ], - [ - "Job Description", - 0.12716859629085603 - ], - [ - "Tech Solutions", - 0.08782397490454795 - ], - [ - "San Francisco", - 0.07719115778822205 - ], - [ - "web application", - 0.03690072848937638 - ], - [ - "experience", - 0.02630939049696386 - ], - [ - "team", - 0.0072399168179765275 - ], - [ - "year", - 0.006650172045860296 - ], - [ - "role", - 0.005752969385077165 - ], - [ - "code", - 0.005550127662632791 - ], - [ - "design", - 0.005254180016905323 - ], - [ - "library", - 0.004910433095184097 - ], - [ - "\u2022familiarity", - 0.003430048956390366 - ], - [ - "frontend", - 0.003422322585261325 - ], - [ - "tool", - 0.003331063486297223 - ], - [ - "USA", - 0.003137490746715329 - ], - [ - "principle", - 0.0028692567567914075 - ], - [ - "candidate", - 0.002853592497377755 - ], - [ - "future", - 0.0028512434085801756 - ] - ], - "name": [ - "Engi" - ], - "phones": [], - "pos_frequencies": { - "ADJ": 33, - "ADP": 36, - "ADV": 3, - "AUX": 10, - "CCONJ": 17, - "DET": 14, - "INTJ": 2, - "NOUN": 88, - "NUM": 4, - "PART": 6, - "PRON": 9, - "PROPN": 51, - "SCONJ": 3, - "SPACE": 36, - "SYM": 1, - "VERB": 22 - }, - "resume_data": "Job Description: Front End Engineer (2 Years of Experience)\nTech Solutions, San Francisco, CA, USA\nAbout Us\nAt Tech Solutions, we are on a mission to build products that solve complex problems and improve people\u2019s\nlives. We are seeking a talented Front End Engineer to join our dynamic team in San Francisco.\nJob Description\nWe are looking for a Front End Engineer with at least 2 years of experience in developing scalable and\nuser-friendly web applications. The successful candidate will be proficient in modern JavaScript frameworks\nand libraries, HTML, CSS, and responsive design principles. This role will contribute significantly to the\ncreation and implementation of user interfaces for our web applications.\nResponsibilities\n\u2022Develop new user-facing features using modern JavaScript frameworks like React.js, Vue.js, or Angu-\nlar.js.\n\u2022Build reusable code and libraries for future use.\n\u2022Ensure the technical feasibility of UI/UX designs.\n\u2022Optimize application for maximum speed and scalability.\n\u2022Assure that all user input is validated before submitting to back-end services.\n\u2022Collaborate with other team members and stakeholders.\nRequirements\n\u20222 years of experience as a Front End Developer or similar role.\n\u2022Proficiency in web markup, including HTML5, CSS3.\n\u2022Knowledge of modern JavaScript programming and experience with libraries like jQuery.\n\u2022Familiarity with modern front-end build pipelines and tools.\n\u2022Experience with popular front-end frameworks (such as React, Vue, or Angular).\n\u2022Familiarity with code versioning tools such as Git.\n\u2022Degree in Computer Science, Engineering, or a related field.\n1 Benefits\n\u2022Competitive salary package.\n\u2022Health, dental, and vision insurance.\n\u2022Retirement savings plan.\n\u2022Professional development opportunities.\n\u2022Flexible work hours.\nTech Solutions is proud to be an equal opportunity employer. We celebrate diversity and are committed\nto creating an inclusive environment for all employees.\nHow to Apply\nTo apply, please submit your resume and a brief explanation of your relevant experience to hiring@techsolutions.com .\n2", - "tri_grams": "[Description Front End, Years of Experience, Tech Solutions San, Solutions San Francisco, Francisco CA USA, mission to build, products that solve, solve complex problems, problems and improve, seeking a talented, talented Front End, Engineer to join, join our dynamic, team in San, years of experience, experience in developing, userfriendly web applications, applications The successful, proficient in modern, modern JavaScript frameworks, libraries HTML CSS, CSS and responsive, responsive design principles, principles This role, role will contribute, creation and implementation, implementation of user, \u2022Develop new userfacing, new userfacing features, features using modern, modern JavaScript frameworks, JavaScript frameworks like, frameworks like Reactjs, like Reactjs Vuejs, Vuejs or Angu, \u2022Build reusable code, code and libraries, libraries for future, \u2022Ensure the technical, feasibility of UI, application for maximum, speed and scalability, input is validated, validated before submitting, submitting to backend, members and stakeholders, years of experience, Developer or similar, \u2022Proficiency in web, web markup including, markup including HTML5, including HTML5 CSS3, \u2022Knowledge of modern, modern JavaScript programming, programming and experience, experience with libraries, libraries like jQuery, \u2022Familiarity with modern, modern frontend build, frontend build pipelines, pipelines and tools, \u2022Experience with popular, popular frontend frameworks, Vue or Angular, \u2022Familiarity with code, code versioning tools, \u2022Degree in Computer, Computer Science Engineering, \u2022Competitive salary package, dental and vision, \u2022Retirement savings plan, \u2022Professional development opportunities, \u2022Flexible work hours, Solutions is proud, equal opportunity employer, employer We celebrate, creating an inclusive, apply please submit, submit your resume]", - "unique_id": "9e508eff-a083-4e2d-8b1e-577cbc8f50fc", - "years": [] -} \ No newline at end of file diff --git a/Data/Processed/JobDescription/JobDescription-job_desc_full_stack_engineer.pdfcccf72e0-5f20-4aa8-8679-d91b720b7247.json b/Data/Processed/JobDescription/JobDescription-job_desc_full_stack_engineer.pdfcccf72e0-5f20-4aa8-8679-d91b720b7247.json deleted file mode 100644 index a0ad8998..00000000 --- a/Data/Processed/JobDescription/JobDescription-job_desc_full_stack_engineer.pdfcccf72e0-5f20-4aa8-8679-d91b720b7247.json +++ /dev/null @@ -1,261 +0,0 @@ -{ - "bi_grams": "[Job Description, Description Senior, Stack Engineer, + Years, Tech Solutions, Solutions San, San Francisco, Tech Solutions, leading technology, technology company, creates innovative, innovative solutions, improve lives, advanced technology, currently seeking, dynamic team, Job Description, Stack Engineer, + years, developing web, web applications, successful candidate, experience working, backend technologies, overseeing projects, production deployment, end website, website architecture, \u2022Designing user, user interactions, web pages, end website, website applications, \u2022Creating servers, \u2022Ensuring crossplatform, crossplatform optimization, mobile phones, finished product, developing APIs, consumer needs, \u2022Staying abreast, web applications, programming languages, Computer Science, similar field, \u20225+, + years, web development, \u2022Strong organizational, project management, management skills, end languages, HTML CSS, serverside languages, Python Ruby, Ruby Java, Java PHP, database technology, MySQL Oracle, \u2022Excellent verbal, verbal communication, communication skills, \u2022Good problemsolving, problemsolving skills, \u2022Competitive salary, salary package, \u2022Health dental, vision insurance, \u2022Retirement savings, savings plan, \u2022Professional development, development opportunities, \u2022Flexible work, work hours, Tech Solutions, equal opportunity, opportunity employer, value diversity, brief explanation, relevant experience]", - "clean_data": "Job Description Senior Full Stack Engineer 5+ Years of\nExperience\nTech Solutions San Francisco CA USA\nAbout Us\nTech Solutions is a leading technology company that creates innovative solutions across a variety of industries\nOur mission is to improve lives through advanced technology We\u2019re currently seeking a Senior Full Stack\nEngineer to join our dynamic team\nJob Description\nWe\u2019re looking for a Senior Full Stack Engineer with 5+ years of experience in developing web applications\nThe successful candidate will have experience working with both frontend and backend technologies and\nwill be capable of overseeing projects from conception to production deployment\nResponsibilities\n\u2022Developing front end website architecture\n\u2022Designing user interactions on web pages\n\u2022Developing back end website applications\n\u2022Creating servers and databases for functionality\n\u2022Ensuring crossplatform optimization for mobile phones\n\u2022Seeing through a project from conception to finished product\n\u2022Designing and developing APIs\n\u2022Meeting both technical and consumer needs\n\u2022Staying abreast of developments in web applications and programming languages\nRequirements\n\u2022Degree in Computer Science or similar field\n\u20225+ years of experience in web development\n\u2022Strong organizational and project management skills\n\u2022Proficiency with fundamental front end languages such as HTML CSS and JavaScript\n\u2022Proficiency with serverside languages such as Python Ruby Java PHP and Net\n\u2022Familiarity with database technology such as MySQL Oracle and MongoDB\n1 \u2022Excellent verbal communication skills\n\u2022Good problemsolving skills\n\u2022Attention to detail\nBenefits\n\u2022Competitive salary package\n\u2022Health dental and vision insurance\n\u2022Retirement savings plan\n\u2022Professional development opportunities\n\u2022Flexible work hours\nTech Solutions is an equal opportunity employer and we value diversity at our company\nHow to Apply\nTo apply please submit your resume and a brief explanation of your relevant experience to \n2", - "emails": [ - "hiring@techsolutions.com" - ], - "entities": [ - "JavaScript", - "\u20225", - "Tech Solutions", - "San Francisco", - "Python Ruby Java PHP", - "Us", - "HTML CSS" - ], - "experience": "Experience \n Tech Solutions San Francisco CA USA \n About Us \n Tech Solutions is a leading technology company that creates innovative solutions across a variety of industries \n Our mission is to improve lives through advanced technology We \u2019re currently seeking a Senior Full Stack \n Engineer to join our dynamic team \n Job Description \n We \u2019re looking for a Senior Full Stack Engineer with 5 + years of experience in developing web applications \n The successful candidate will have experience working with both frontend and backend technologies and \n will be capable of overseeing projects from conception to production deployment \n Responsibilities \n \u2022Developing front end website architecture \n \u2022Designing user interactions on web pages \n \u2022Developing back end website applications \n \u2022Creating servers and databases for functionality \n \u2022Ensuring crossplatform optimization for mobile phones \n \u2022Seeing through a project from conception to finished product \n \u2022Designing and developing APIs \n \u2022Meeting both technical and consumer needs \n \u2022Staying abreast of developments in web applications and programming languages \n Requirements \n \u2022Degree in Computer Science or similar field \n \u20225 + years of experience in web development \n \u2022Strong organizational and project management skills \n \u2022Proficiency with fundamental front end languages such as HTML CSS and JavaScript \n \u2022Proficiency with serverside languages such as Python Ruby Java PHP and Net \n \u2022Familiarity with database technology such as MySQL Oracle and MongoDB \n 1 \u2022Excellent verbal communication skills \n \u2022Good problemsolving skills \n \u2022Attention to detail \n Benefits \n \u2022Competitive salary package \n \u2022Health dental and vision insurance \n \u2022Retirement savings plan \n \u2022Professional development opportunities \n \u2022Flexible work hours \n Tech Solutions is an equal opportunity employer and we value diversity at our company \n How to Apply \n To apply please submit your resume and a brief explanation of your relevant experience to \n 2", - "extracted_keywords": [ - "Job", - "Description", - "Senior", - "Full", - "Stack", - "Engineer", - "Years", - "Experience", - "Tech", - "Solutions", - "San", - "Francisco", - "CA", - "USA", - "Us", - "Tech", - "Solutions", - "technology", - "company", - "solutions", - "variety", - "industries", - "mission", - "lives", - "technology", - "Senior", - "Full", - "Stack", - "Engineer", - "team", - "Job", - "Description", - "Senior", - "Full", - "Stack", - "Engineer", - "years", - "experience", - "web", - "applications", - "candidate", - "experience", - "frontend", - "technologies", - "projects", - "conception", - "production", - "deployment", - "Responsibilities", - "end", - "website", - "architecture", - "\u2022Designing", - "user", - "interactions", - "web", - "pages", - "end", - "website", - "applications", - "servers", - "databases", - "functionality", - "crossplatform", - "optimization", - "phones", - "project", - "conception", - "product", - "\u2022Designing", - "APIs", - "consumer", - "needs", - "\u2022Staying", - "developments", - "web", - "applications", - "programming", - "languages", - "Requirements", - "Computer", - "Science", - "field", - "\u20225", - "years", - "experience", - "web", - "development", - "\u2022Strong", - "project", - "management", - "skills", - "\u2022Proficiency", - "end", - "languages", - "HTML", - "CSS", - "JavaScript", - "\u2022Proficiency", - "serverside", - "languages", - "Python", - "Ruby", - "Java", - "PHP", - "Net", - "\u2022Familiarity", - "database", - "technology", - "MySQL", - "Oracle", - "MongoDB", - "communication", - "skills", - "\u2022Attention", - "Benefits", - "salary", - "package", - "\u2022Health", - "vision", - "insurance", - "savings", - "plan", - "development", - "opportunities", - "work", - "hours", - "Tech", - "Solutions", - "opportunity", - "employer", - "diversity", - "company", - "resume", - "explanation", - "experience" - ], - "keyterms": [ - [ - "Senior Full stack", - 0.36734547073779045 - ], - [ - "Job Description", - 0.1137969836501915 - ], - [ - "Tech Solutions", - 0.08785209480327773 - ], - [ - "experience", - 0.0402594489684917 - ], - [ - "web application", - 0.031091464413874694 - ], - [ - "technology", - 0.026103226555550127 - ], - [ - "end website", - 0.02603264785474888 - ], - [ - "year", - 0.02280185463098348 - ], - [ - "Engineer", - 0.015158334006474023 - ], - [ - "company", - 0.01299126818329351 - ], - [ - "project", - 0.008025371467830188 - ], - [ - "development", - 0.006199971245814812 - ], - [ - "database", - 0.005431870607290941 - ], - [ - "conception", - 0.005159537892195308 - ], - [ - "language", - 0.004950592066886817 - ], - [ - "solution", - 0.004909896555990972 - ], - [ - "innovative", - 0.004862122886957498 - ], - [ - "Francisco", - 0.004742057116465734 - ], - [ - "variety", - 0.004727628158636155 - ], - [ - "San", - 0.004700569886380095 - ] - ], - "name": [], - "phones": [], - "pos_frequencies": { - "ADJ": 28, - "ADP": 33, - "ADV": 1, - "AUX": 8, - "CCONJ": 16, - "DET": 8, - "INTJ": 2, - "NOUN": 99, - "NUM": 5, - "PART": 5, - "PRON": 9, - "PROPN": 38, - "SCONJ": 1, - "SPACE": 40, - "SYM": 2, - "VERB": 25 - }, - "resume_data": "Job Description: Senior Full Stack Engineer (5+ Years of\nExperience)\nTech Solutions, San Francisco, CA, USA\nAbout Us\nTech Solutions is a leading technology company that creates innovative solutions across a variety of industries.\nOur mission is to improve lives through advanced technology. We\u2019re currently seeking a Senior Full Stack\nEngineer to join our dynamic team.\nJob Description\nWe\u2019re looking for a Senior Full Stack Engineer with 5+ years of experience in developing web applications.\nThe successful candidate will have experience working with both front-end and back-end technologies, and\nwill be capable of overseeing projects from conception to production deployment.\nResponsibilities\n\u2022Developing front end website architecture.\n\u2022Designing user interactions on web pages.\n\u2022Developing back end website applications.\n\u2022Creating servers and databases for functionality.\n\u2022Ensuring cross-platform optimization for mobile phones.\n\u2022Seeing through a project from conception to finished product.\n\u2022Designing and developing APIs.\n\u2022Meeting both technical and consumer needs.\n\u2022Staying abreast of developments in web applications and programming languages.\nRequirements\n\u2022Degree in Computer Science or similar field.\n\u20225+ years of experience in web development.\n\u2022Strong organizational and project management skills.\n\u2022Proficiency with fundamental front end languages such as HTML, CSS, and JavaScript.\n\u2022Proficiency with server-side languages such as Python, Ruby, Java, PHP, and .Net.\n\u2022Familiarity with database technology such as MySQL, Oracle, and MongoDB.\n1 \u2022Excellent verbal communication skills.\n\u2022Good problem-solving skills.\n\u2022Attention to detail.\nBenefits\n\u2022Competitive salary package.\n\u2022Health, dental, and vision insurance.\n\u2022Retirement savings plan.\n\u2022Professional development opportunities.\n\u2022Flexible work hours.\nTech Solutions is an equal opportunity employer and we value diversity at our company.\nHow to Apply\nTo apply, please submit your resume and a brief explanation of your relevant experience to hiring@techsolutions.com .\n2", - "tri_grams": "[Job Description Senior, Senior Full Stack, Tech Solutions San, Solutions San Francisco, Francisco CA USA, leading technology company, company that creates, creates innovative solutions, variety of industries, lives through advanced, seeking a Senior, Senior Full Stack, Engineer to join, join our dynamic, Senior Full Stack, years of experience, experience in developing, developing web applications, frontend and backend, capable of overseeing, projects from conception, conception to production, \u2022Developing front end, end website architecture, \u2022Designing user interactions, interactions on web, \u2022Developing back end, end website applications, servers and databases, databases for functionality, \u2022Ensuring crossplatform optimization, optimization for mobile, project from conception, conception to finished, \u2022Designing and developing, \u2022Meeting both technical, technical and consumer, abreast of developments, developments in web, applications and programming, \u2022Degree in Computer, Science or similar, \u20225+ years, years of experience, experience in web, organizational and project, project management skills, \u2022Proficiency with fundamental, fundamental front end, CSS and JavaScript, \u2022Proficiency with serverside, Python Ruby Java, Ruby Java PHP, PHP and Net, \u2022Familiarity with database, Oracle and MongoDB, \u2022Excellent verbal communication, verbal communication skills, \u2022Good problemsolving skills, \u2022Attention to detail, \u2022Competitive salary package, dental and vision, \u2022Retirement savings plan, \u2022Professional development opportunities, \u2022Flexible work hours, equal opportunity employer, apply please submit, submit your resume]", - "unique_id": "cccf72e0-5f20-4aa8-8679-d91b720b7247", - "years": [] -} \ No newline at end of file diff --git a/Data/Processed/JobDescription/JobDescription-job_desc_java_developer.pdf1fb6435c-531a-4f04-84c2-c4e73e1f1a3f.json b/Data/Processed/JobDescription/JobDescription-job_desc_java_developer.pdf1fb6435c-531a-4f04-84c2-c4e73e1f1a3f.json deleted file mode 100644 index fb0cb9f3..00000000 --- a/Data/Processed/JobDescription/JobDescription-job_desc_java_developer.pdf1fb6435c-531a-4f04-84c2-c4e73e1f1a3f.json +++ /dev/null @@ -1,283 +0,0 @@ -{ - "bi_grams": "[Job Description, Description Java, Java Developer, Tech Solutions, Solutions San, San Francisco, Tech Solutions, solve complex, complex problems, forwardthinking tech, tech company, company specializing, custom software, software solutions, experienced Java, Java Developer, Job Description, skilled Java, Java Developer, building highperforming, highperforming scal, able enterprisegrade, enterprisegrade applications, talented software, software team, include managing, managing Java, Java EE, EE application, application development, providing expertise, software development, development lifecycle, \u2022Designing implementing, maintaining Java, Java applications, latency required, missioncritical systems, \u2022Delivering high, high availability, development lifecycle, \u2022Writing welldesigned, welldesigned efficient, testable code, \u2022Conducting software, software analysis, analysis programming, programming testing, \u2022Ensuring designs, designs comply, producing releases, software components, \u2022Supporting continuous, continuous improvement, investigating alternatives, architectural review, MS degree, Computer Science, Science Engineering, related subject, \u2022Proven handson, handson Software, Software Development, Development experience, \u2022Proven working, working experience, Java development, \u2022Handson experience, developing applications, Java EE, EE platforms, \u2022ObjectOriented Analysis, common design, design patterns, \u2022Profound insight, JEE internals, internals Classloading, Classloading Memory, Memory Management, Management Transaction, Transaction man, agement etc, \u2022Excellent knowledge, Relational Databases, Databases SQL, ORM technologies, technologies JPA2, JPA2 Hibernate, developing web, web applications, popular web, web framework, framework JSF, JSF Wicket, GWT Spring, Spring MVC, testdriven development, \u2022Competitive salary, salary package, \u2022Health dental, vision insurance, \u2022Retirement savings, savings plan, \u2022Professional development, development opportunities, \u2022Flexible work, work hours, Tech Solutions, equal opportunity, opportunity employer, celebrate diversity, inclusive environment, brief explanation, relevant experience]", - "clean_data": "Job Description Java Developer 3 Years of Experience\nTech Solutions San Francisco CA USA\nAbout Us\nAt Tech Solutions we believe in the power of technology to solve complex problems We are a dynamic\nforwardthinking tech company specializing in custom software solutions for various industries We are\nseeking a talented and experienced Java Developer to join our team\nJob Description\nWe are seeking a skilled Java Developer with at least 3 years of experience in building highperforming scal\nable enterprisegrade applications You will be part of a talented software team that works on missioncritical\napplications Your roles and responsibilities will include managing Java/Java EE application development\nwhile providing expertise in the full software development lifecycle\nResponsibilities\n\u2022Designing implementing and maintaining Java applications that are often highvolume and low\nlatency required for missioncritical systems\n\u2022Delivering high availability and performance\n\u2022Contributing to all phases of the development lifecycle\n\u2022Writing welldesigned efficient and testable code\n\u2022Conducting software analysis programming testing and debugging\n\u2022Ensuring designs comply with specifications\n\u2022Preparing and producing releases of software components\n\u2022Supporting continuous improvement by investigating alternatives and technologies and presenting these\nfor architectural review\nRequirements\n\u2022BS/MS degree in Computer Science Engineering or a related subject\n\u2022Proven handson Software Development experience\n\u2022Proven working experience in Java development\n\u2022Handson experience in designing and developing applications using Java EE platforms\n\u2022ObjectOriented Analysis and design using common design patterns\n\u2022Profound insight of Java and JEE internals Classloading Memory Management Transaction man\nagement etc\n1 \u2022Excellent knowledge of Relational Databases SQL and ORM technologies JPA2 Hibernate\n\u2022Experience in developing web applications using at least one popular web framework JSF Wicket\nGWT Spring MVC\n\u2022Experience with testdriven development\nBenefits\n\u2022Competitive salary package\n\u2022Health dental and vision insurance\n\u2022Retirement savings plan\n\u2022Professional development opportunities\n\u2022Flexible work hours\nTech Solutions is proud to be an equal opportunity employer We celebrate diversity and are committed\nto creating an inclusive environment for all employees\nHow to Apply\nTo apply please submit your resume and a brief explanation of your relevant experience to \n2", - "emails": [ - "hiring@techsolutions.com" - ], - "entities": [ - "GWT Spring", - "Tech Solutions", - "Computer Science Engineering", - "San Francisco", - "ORM", - "MVC", - "Classloading Memory Management Transaction", - "JSF Wicket", - "Us", - "\u2022Proven handson Software Development" - ], - "experience": "Experience \n Tech Solutions San Francisco CA USA \n About Us \n At Tech Solutions we believe in the power of technology to solve complex problems We are a dynamic \n forwardthinking tech company specializing in custom software solutions for various industries We are \n seeking a talented and experienced Java Developer to join our team \n Job Description \n We are seeking a skilled Java Developer with at least 3 years of experience in building highperforming scal \n able enterprisegrade applications You will be part of a talented software team that works on missioncritical \n applications Your roles and responsibilities will include managing Java / Java EE application development \n while providing expertise in the full software development lifecycle \n Responsibilities \n \u2022Designing implementing and maintaining Java applications that are often highvolume and low \n latency required for missioncritical systems \n \u2022Delivering high availability and performance \n \u2022Contributing to all phases of the development lifecycle \n \u2022Writing welldesigned efficient and testable code \n \u2022Conducting software analysis programming testing and debugging \n \u2022Ensuring designs comply with specifications \n \u2022Preparing and producing releases of software components \n \u2022Supporting continuous improvement by investigating alternatives and technologies and presenting these \n for architectural review \n Requirements \n \u2022BS / MS degree in Computer Science Engineering or a related subject \n \u2022Proven handson Software Development experience \n \u2022Proven working experience in Java development \n \u2022Handson experience in designing and developing applications using Java EE platforms \n \u2022ObjectOriented Analysis and design using common design patterns \n \u2022Profound insight of Java and JEE internals Classloading Memory Management Transaction man \n agement etc \n 1 \u2022Excellent knowledge of Relational Databases SQL and ORM technologies JPA2 Hibernate \n \u2022Experience in developing web applications using at least one popular web framework JSF Wicket \n GWT Spring MVC \n \u2022Experience with testdriven development \n Benefits \n \u2022Competitive salary package \n \u2022Health dental and vision insurance \n \u2022Retirement savings plan \n \u2022Professional development opportunities \n \u2022Flexible work hours \n Tech Solutions is proud to be an equal opportunity employer We celebrate diversity and are committed \n to creating an inclusive environment for all employees \n How to Apply \n To apply please submit your resume and a brief explanation of your relevant experience to \n 2", - "extracted_keywords": [ - "Job", - "Description", - "Java", - "Developer", - "Years", - "Experience", - "Tech", - "Solutions", - "San", - "Francisco", - "CA", - "USA", - "Tech", - "Solutions", - "power", - "technology", - "problems", - "tech", - "company", - "custom", - "software", - "solutions", - "industries", - "Java", - "Developer", - "team", - "Job", - "Description", - "Java", - "Developer", - "years", - "experience", - "enterprisegrade", - "applications", - "part", - "software", - "team", - "applications", - "roles", - "responsibilities", - "Java", - "Java", - "EE", - "application", - "development", - "expertise", - "software", - "development", - "lifecycle", - "Responsibilities", - "Java", - "applications", - "highvolume", - "latency", - "systems", - "availability", - "performance", - "phases", - "development", - "lifecycle", - "\u2022Writing", - "code", - "software", - "analysis", - "programming", - "testing", - "\u2022Ensuring", - "designs", - "specifications", - "\u2022Preparing", - "releases", - "software", - "components", - "improvement", - "alternatives", - "technologies", - "review", - "Requirements", - "\u2022BS", - "MS", - "degree", - "Computer", - "Science", - "Engineering", - "subject", - "handson", - "Software", - "Development", - "experience", - "experience", - "Java", - "development", - "\u2022Handson", - "experience", - "applications", - "Java", - "EE", - "platforms", - "Analysis", - "design", - "design", - "patterns", - "insight", - "Java", - "JEE", - "internals", - "Classloading", - "Memory", - "Management", - "Transaction", - "man", - "agement", - "knowledge", - "Relational", - "Databases", - "SQL", - "ORM", - "technologies", - "JPA2", - "Hibernate", - "\u2022Experience", - "web", - "applications", - "web", - "framework", - "JSF", - "Wicket", - "GWT", - "Spring", - "MVC", - "\u2022Experience", - "testdriven", - "development", - "Benefits", - "salary", - "package", - "\u2022Health", - "vision", - "insurance", - "savings", - "plan", - "development", - "opportunities", - "work", - "hours", - "Tech", - "Solutions", - "opportunity", - "employer", - "diversity", - "environment", - "employees", - "resume", - "explanation", - "experience" - ], - "keyterms": [ - [ - "Job Description", - 0.19752577775414384 - ], - [ - "Java Developer", - 0.17768673477735694 - ], - [ - "Tech Solutions", - 0.1625520418906361 - ], - [ - "development lifecycle", - 0.04326534361760704 - ], - [ - "Java EE", - 0.03633403178723749 - ], - [ - "experience", - 0.028543152722860825 - ], - [ - "application", - 0.0215287649967145 - ], - [ - "software", - 0.018694891435522134 - ], - [ - "technology", - 0.01021144169169241 - ], - [ - "talented", - 0.0070539496077944245 - ], - [ - "team", - 0.0067607132681006335 - ], - [ - "design", - 0.0054077035745175766 - ], - [ - "missioncritical", - 0.0052934339438335315 - ], - [ - "responsibility", - 0.005155750713317631 - ], - [ - "analysis", - 0.004650613996899872 - ], - [ - "Francisco", - 0.0043704900410092675 - ], - [ - "San", - 0.00432692619225915 - ], - [ - "USA", - 0.0042781236165988805 - ], - [ - "solution", - 0.004146827223158856 - ], - [ - "industry", - 0.003994745329659428 - ] - ], - "name": [], - "phones": [], - "pos_frequencies": { - "ADJ": 35, - "ADP": 31, - "ADV": 5, - "AUX": 10, - "CCONJ": 18, - "DET": 13, - "INTJ": 1, - "NOUN": 102, - "NUM": 5, - "PART": 5, - "PRON": 13, - "PROPN": 52, - "SCONJ": 3, - "SPACE": 44, - "SYM": 2, - "VERB": 43 - }, - "resume_data": "Job Description: Java Developer (3 Years of Experience)\nTech Solutions, San Francisco, CA, USA\nAbout Us\nAt Tech Solutions, we believe in the power of technology to solve complex problems. We are a dynamic,\nforward-thinking tech company specializing in custom software solutions for various industries. We are\nseeking a talented and experienced Java Developer to join our team.\nJob Description\nWe are seeking a skilled Java Developer with at least 3 years of experience in building high-performing, scal-\nable, enterprise-grade applications. You will be part of a talented software team that works on mission-critical\napplications. Your roles and responsibilities will include managing Java/Java EE application development\nwhile providing expertise in the full software development lifecycle.\nResponsibilities\n\u2022Designing, implementing, and maintaining Java applications that are often high-volume and low-\nlatency, required for mission-critical systems.\n\u2022Delivering high availability and performance.\n\u2022Contributing to all phases of the development lifecycle.\n\u2022Writing well-designed, efficient, and testable code.\n\u2022Conducting software analysis, programming, testing, and debugging.\n\u2022Ensuring designs comply with specifications.\n\u2022Preparing and producing releases of software components.\n\u2022Supporting continuous improvement by investigating alternatives and technologies and presenting these\nfor architectural review.\nRequirements\n\u2022BS/MS degree in Computer Science, Engineering, or a related subject.\n\u2022Proven hands-on Software Development experience.\n\u2022Proven working experience in Java development.\n\u2022Hands-on experience in designing and developing applications using Java EE platforms.\n\u2022Object-Oriented Analysis and design using common design patterns.\n\u2022Profound insight of Java and JEE internals (Classloading, Memory Management, Transaction man-\nagement, etc.)\n1 \u2022Excellent knowledge of Relational Databases, SQL, and ORM technologies (JPA2, Hibernate).\n\u2022Experience in developing web applications using at least one popular web framework (JSF, Wicket,\nGWT, Spring MVC).\n\u2022Experience with test-driven development.\nBenefits\n\u2022Competitive salary package.\n\u2022Health, dental, and vision insurance.\n\u2022Retirement savings plan.\n\u2022Professional development opportunities.\n\u2022Flexible work hours.\nTech Solutions is proud to be an equal opportunity employer. We celebrate diversity and are committed\nto creating an inclusive environment for all employees.\nHow to Apply\nTo apply, please submit your resume and a brief explanation of your relevant experience to hiring@techsolutions.com .\n2", - "tri_grams": "[Job Description Java, Description Java Developer, Years of Experience, Tech Solutions San, Solutions San Francisco, Francisco CA USA, Solutions we believe, power of technology, technology to solve, solve complex problems, forwardthinking tech company, tech company specializing, specializing in custom, custom software solutions, seeking a talented, talented and experienced, experienced Java Developer, Developer to join, join our team, seeking a skilled, skilled Java Developer, years of experience, experience in building, building highperforming scal, able enterprisegrade applications, talented software team, team that works, works on missioncritical, applications Your roles, roles and responsibilities, responsibilities will include, include managing Java, Java EE application, EE application development, software development lifecycle, implementing and maintaining, maintaining Java applications, highvolume and low, required for missioncritical, \u2022Delivering high availability, availability and performance, \u2022Writing welldesigned efficient, efficient and testable, \u2022Conducting software analysis, software analysis programming, analysis programming testing, testing and debugging, \u2022Ensuring designs comply, comply with specifications, \u2022Preparing and producing, releases of software, \u2022Supporting continuous improvement, improvement by investigating, alternatives and technologies, technologies and presenting, degree in Computer, Computer Science Engineering, \u2022Proven handson Software, handson Software Development, Software Development experience, \u2022Proven working experience, experience in Java, experience in designing, designing and developing, applications using Java, Java EE platforms, Analysis and design, design using common, common design patterns, insight of Java, Java and JEE, JEE internals Classloading, internals Classloading Memory, Classloading Memory Management, Memory Management Transaction, Management Transaction man, knowledge of Relational, Relational Databases SQL, SQL and ORM, ORM technologies JPA2, technologies JPA2 Hibernate, \u2022Experience in developing, developing web applications, popular web framework, web framework JSF, framework JSF Wicket, GWT Spring MVC, \u2022Experience with testdriven, \u2022Competitive salary package, dental and vision, \u2022Retirement savings plan, \u2022Professional development opportunities, \u2022Flexible work hours, Solutions is proud, equal opportunity employer, employer We celebrate, creating an inclusive, apply please submit, submit your resume]", - "unique_id": "1fb6435c-531a-4f04-84c2-c4e73e1f1a3f", - "years": [] -} \ No newline at end of file diff --git a/Data/Processed/JobDescription/JobDescription-job_desc_product_manager.pdf0b613898-1183-49c5-ad07-e03bd9af72e5.json b/Data/Processed/JobDescription/JobDescription-job_desc_product_manager.pdf0b613898-1183-49c5-ad07-e03bd9af72e5.json deleted file mode 100644 index c8ba755a..00000000 --- a/Data/Processed/JobDescription/JobDescription-job_desc_product_manager.pdf0b613898-1183-49c5-ad07-e03bd9af72e5.json +++ /dev/null @@ -1,247 +0,0 @@ -{ - "bi_grams": "[Job Description, Description Product, Product Manager, + Years, Tech Solutions, Solutions San, San Francisco, Tech Solutions, global leader, technology industry, industry specializing, software products, currently looking, seasoned Product, Product Manager, dynamic team, Job Description, Product Manager, important organizational, organizational role, feature definition, product line, product strategy, \u2022Deliver MRDs, prioritized features, corresponding justification, assess partnerships, licensing opportunities, \u2022Run beta, pilot programs, earlystage products, core positioning, \u2022Perform product, product demos, \u2022Set pricing, meet revenue, profitability goals, \u202210+, + years, product management, \u2022Demonstrated success, success defining, launching excellent, excellent products, \u2022Excellent written, verbal communication, communication skills, \u2022Technical background, software development, \u2022Excellent teamwork, teamwork skills, \u2022Proven ability, influence crossfunctional, crossfunctional teams, formal authority, degree MBA, MBA preferred, \u2022Competitive salary, salary package, \u2022Health dental, vision insurance, \u2022Retirement savings, savings plan, \u2022Professional development, development opportunities, \u2022Flexible work, work hours, Tech Solutions, equal opportunity, opportunity employer, celebrate diversity, inclusive environment, brief explanation, relevant experience]", - "clean_data": "Job Description Product Manager 10+ Years of Experience\nTech Solutions San Francisco CA USA\nAbout Us\nTech Solutions is a global leader in the technology industry specializing in the development of cuttingedge\nsoftware products We\u2019re currently looking for a seasoned Product Manager with over 10 years of experience\nto join our dynamic team\nJob Description\nThe Product Manager will be responsible for guiding the success of a product and leading the crossfunctional\nteam that is responsible for improving it This is an important organizational role that sets the strategy\nroadmap and feature definition for a product or product line\nResponsibilities\n\u2022Define the product strategy and roadmap\n\u2022Deliver MRDs and PRDs with prioritized features and corresponding justification\n\u2022Work with external third parties to assess partnerships and licensing opportunities\n\u2022Run beta and pilot programs with earlystage products and samples\n\u2022Be an expert with respect to the competition\n\u2022Act as a leader within the company\n\u2022Develop the core positioning and messaging for the product\n\u2022Perform product demos to customers\n\u2022Set pricing to meet revenue and profitability goals\nRequirements\n\u202210+ years of experience in product management\n\u2022Demonstrated success defining and launching excellent products\n\u2022Excellent written and verbal communication skills\n\u2022Technical background with experience in software development\n\u2022Excellent teamwork skills\n\u2022Proven ability to influence crossfunctional teams without formal authority\n\u2022Must be able to travel 20\n\u2022Bachelor\u2019s degree MBA preferred\n1 Benefits\n\u2022Competitive salary package\n\u2022Health dental and vision insurance\n\u2022Retirement savings plan\n\u2022Professional development opportunities\n\u2022Flexible work hours\nTech Solutions is an equal opportunity employer We celebrate diversity and are committed to creating\nan inclusive environment for all employees\nHow to Apply\nTo apply please submit your resume and a brief explanation of your relevant experience to \n2", - "emails": [ - "hiring@techsolutions.com" - ], - "entities": [ - "Tech Solutions", - "San Francisco", - "Us", - "Requirements" - ], - "experience": "Experience \n Tech Solutions San Francisco CA USA \n About Us \n Tech Solutions is a global leader in the technology industry specializing in the development of cuttingedge \n software products We \u2019re currently looking for a seasoned Product Manager with over 10 years of experience \n to join our dynamic team \n Job Description \n The Product Manager will be responsible for guiding the success of a product and leading the crossfunctional \n team that is responsible for improving it This is an important organizational role that sets the strategy \n roadmap and feature definition for a product or product line \n Responsibilities \n \u2022Define the product strategy and roadmap \n \u2022Deliver MRDs and PRDs with prioritized features and corresponding justification \n \u2022Work with external third parties to assess partnerships and licensing opportunities \n \u2022Run beta and pilot programs with earlystage products and samples \n \u2022Be an expert with respect to the competition \n \u2022Act as a leader within the company \n \u2022Develop the core positioning and messaging for the product \n \u2022Perform product demos to customers \n \u2022Set pricing to meet revenue and profitability goals \n Requirements \n \u202210 + years of experience in product management \n \u2022Demonstrated success defining and launching excellent products \n \u2022Excellent written and verbal communication skills \n \u2022Technical background with experience in software development \n \u2022Excellent teamwork skills \n \u2022Proven ability to influence crossfunctional teams without formal authority \n \u2022Must be able to travel 20 \n \u2022Bachelor \u2019s degree MBA preferred \n 1 Benefits \n \u2022Competitive salary package \n \u2022Health dental and vision insurance \n \u2022Retirement savings plan \n \u2022Professional development opportunities \n \u2022Flexible work hours \n Tech Solutions is an equal opportunity employer We celebrate diversity and are committed to creating \n an inclusive environment for all employees \n How to Apply \n To apply please submit your resume and a brief explanation of your relevant experience to \n 2", - "extracted_keywords": [ - "Job", - "Description", - "Product", - "Manager", - "Years", - "Experience", - "Tech", - "Solutions", - "San", - "Francisco", - "CA", - "USA", - "Us", - "Tech", - "Solutions", - "leader", - "technology", - "industry", - "development", - "cuttingedge", - "software", - "products", - "Product", - "Manager", - "years", - "experience", - "team", - "Job", - "Description", - "Product", - "Manager", - "success", - "product", - "team", - "role", - "strategy", - "roadmap", - "definition", - "product", - "product", - "line", - "Responsibilities", - "\u2022Define", - "product", - "strategy", - "roadmap", - "MRDs", - "PRDs", - "features", - "justification", - "\u2022Work", - "parties", - "partnerships", - "licensing", - "opportunities", - "beta", - "pilot", - "programs", - "earlystage", - "products", - "samples", - "expert", - "respect", - "competition", - "\u2022Act", - "leader", - "company", - "\u2022Develop", - "core", - "positioning", - "product", - "\u2022Perform", - "product", - "demos", - "customers", - "\u2022Set", - "pricing", - "revenue", - "profitability", - "goals", - "Requirements", - "\u202210", - "years", - "experience", - "product", - "management", - "success", - "products", - "communication", - "background", - "experience", - "software", - "development", - "\u2022Excellent", - "teamwork", - "skills", - "ability", - "teams", - "authority", - "\u2022Must", - "\u2022Bachelor", - "degree", - "MBA", - "Benefits", - "salary", - "package", - "\u2022Health", - "vision", - "insurance", - "savings", - "plan", - "development", - "opportunities", - "work", - "hours", - "Tech", - "Solutions", - "opportunity", - "employer", - "diversity", - "environment", - "employees", - "resume", - "explanation", - "experience" - ], - "keyterms": [ - [ - "Product Manager", - 0.2610812469026535 - ], - [ - "Job Description", - 0.21213371354900168 - ], - [ - "Tech Solutions", - 0.12938404019315197 - ], - [ - "experience", - 0.03987149457061917 - ], - [ - "product", - 0.037803890917222985 - ], - [ - "year", - 0.023390376629897593 - ], - [ - "development", - 0.01926741786356023 - ], - [ - "software", - 0.010877641839805458 - ], - [ - "leader", - 0.01072589628317647 - ], - [ - "team", - 0.00896251970410434 - ], - [ - "crossfunctional", - 0.007394518043374325 - ], - [ - "success", - 0.007335957307008526 - ], - [ - "opportunity", - 0.0071798132340313734 - ], - [ - "responsible", - 0.005862432001402156 - ], - [ - "strategy", - 0.004991460820656987 - ], - [ - "roadmap", - 0.0048401824745963875 - ], - [ - "cuttingedge", - 0.004068297634876913 - ], - [ - "dynamic", - 0.0040604308400752276 - ], - [ - "Francisco", - 0.0040586982182202495 - ], - [ - "San", - 0.004034781797427738 - ] - ], - "name": [], - "phones": [], - "pos_frequencies": { - "ADJ": 29, - "ADP": 32, - "ADV": 2, - "AUX": 9, - "CCONJ": 16, - "DET": 22, - "INTJ": 1, - "NOUN": 94, - "NUM": 6, - "PART": 8, - "PRON": 9, - "PROPN": 29, - "SCONJ": 2, - "SPACE": 39, - "SYM": 2, - "VERB": 26 - }, - "resume_data": "Job Description: Product Manager (10+ Years of Experience)\nTech Solutions, San Francisco, CA, USA\nAbout Us\nTech Solutions is a global leader in the technology industry, specializing in the development of cutting-edge\nsoftware products. We\u2019re currently looking for a seasoned Product Manager with over 10 years of experience\nto join our dynamic team.\nJob Description\nThe Product Manager will be responsible for guiding the success of a product and leading the cross-functional\nteam that is responsible for improving it. This is an important organizational role that sets the strategy,\nroadmap, and feature definition for a product or product line.\nResponsibilities\n\u2022Define the product strategy and roadmap.\n\u2022Deliver MRDs and PRDs with prioritized features and corresponding justification.\n\u2022Work with external third parties to assess partnerships and licensing opportunities.\n\u2022Run beta and pilot programs with early-stage products and samples.\n\u2022Be an expert with respect to the competition.\n\u2022Act as a leader within the company.\n\u2022Develop the core positioning and messaging for the product.\n\u2022Perform product demos to customers.\n\u2022Set pricing to meet revenue and profitability goals.\nRequirements\n\u202210+ years of experience in product management.\n\u2022Demonstrated success defining and launching excellent products.\n\u2022Excellent written and verbal communication skills.\n\u2022Technical background, with experience in software development.\n\u2022Excellent teamwork skills.\n\u2022Proven ability to influence cross-functional teams without formal authority.\n\u2022Must be able to travel 20\n\u2022Bachelor\u2019s degree (MBA preferred).\n1 Benefits\n\u2022Competitive salary package.\n\u2022Health, dental, and vision insurance.\n\u2022Retirement savings plan.\n\u2022Professional development opportunities.\n\u2022Flexible work hours.\nTech Solutions is an equal opportunity employer. We celebrate diversity and are committed to creating\nan inclusive environment for all employees.\nHow to Apply\nTo apply, please submit your resume and a brief explanation of your relevant experience to hiring@techsolutions.com .\n2", - "tri_grams": "[Job Description Product, Description Product Manager, Years of Experience, Tech Solutions San, Solutions San Francisco, Francisco CA USA, technology industry specializing, development of cuttingedge, seasoned Product Manager, years of experience, join our dynamic, responsible for guiding, guiding the success, product and leading, leading the crossfunctional, responsible for improving, important organizational role, role that sets, sets the strategy, roadmap and feature, product or product, \u2022Define the product, strategy and roadmap, MRDs and PRDs, PRDs with prioritized, features and corresponding, \u2022Work with external, parties to assess, partnerships and licensing, beta and pilot, programs with earlystage, products and samples, \u2022Be an expert, expert with respect, \u2022Develop the core, positioning and messaging, \u2022Perform product demos, demos to customers, pricing to meet, revenue and profitability, \u202210+ years, years of experience, experience in product, \u2022Demonstrated success defining, defining and launching, launching excellent products, written and verbal, verbal communication skills, background with experience, experience in software, \u2022Excellent teamwork skills, ability to influence, influence crossfunctional teams, teams without formal, \u2022Must be able, able to travel, \u2022Bachelor\u2019s degree, degree MBA preferred, \u2022Competitive salary package, dental and vision, \u2022Retirement savings plan, \u2022Professional development opportunities, \u2022Flexible work hours, equal opportunity employer, employer We celebrate, committed to creating, apply please submit, submit your resume]", - "unique_id": "0b613898-1183-49c5-ad07-e03bd9af72e5", - "years": [] -} \ No newline at end of file diff --git a/Data/Processed/Resumes/Resume-alfred_pennyworth_pm.pdfb36fac00-1d60-49ba-a9e4-33477c928e98.json b/Data/Processed/Resumes/Resume-alfred_pennyworth_pm.pdfb36fac00-1d60-49ba-a9e4-33477c928e98.json deleted file mode 100644 index 84952fe6..00000000 --- a/Data/Processed/Resumes/Resume-alfred_pennyworth_pm.pdfb36fac00-1d60-49ba-a9e4-33477c928e98.json +++ /dev/null @@ -1,324 +0,0 @@ -{ - "bi_grams": "[Alfred Pennyworth, Product ManagerSilicon, ManagerSilicon Valley, \u2642\u00b6obilealt, /envel\u2322, Professional Summary, Seasoned Product, Product Manager, software development, management having, having worked, FAANG companies, companies Exceptional, Exceptional leadership, leadership skills, skills strategic, track record, managing products, market success, Product management, management Agile, Agile methodologies, methodologies Leadership, Leadership Communication, Communication Project, management User, User Experience, Experience Design, Design Market, Market Research, Research Data, Data Analysis, Analysis Java, Python JavaScript, JavaScript HTML, CSS SQL, SQL AWS, PresentProduct Manager, Manager Google, Mountain View, Leading crossfunctional, crossfunctional teams, design develop, launch innovative, innovative products, products Devel, oping product, product strategies, making datadriven, datadriven decisions, improve user, user experience, meet business, business goals, Software Development, Development Engineer, Engineer III, III Amazon, Seattle WA, WA USA, building scalable, highperforming ecommerce, ecommerce applications, Successfully delivered, delivered multiple, multiple projects, stipulated time, Software Development, Development Engineer, Engineer II, II Apple, implemented software, software components, Apple services, services Optimized, improved code, code quality, thorough testing, Software Development, Development Engineer, Los Gatos, user interface, Netflix web, web application, application Worked, Worked closely, product managers, optimal user, user experience, Software Development, Development Engineer, Menlo Park, key role, early Facebook, Facebook features, features Implemented, Implemented scalable, backend services, Business Administration, Administration Stanford, Stanford University, Computer Science, Science Massachusetts, Massachusetts Institute, Cambridge MA, MA USA, Computer Science, Science University, California Berkeley, PresentPersonal Project, Project Home, Home Automation, Automation System, smart home, home automation, automation system, Raspberry Pi, home appliances, appliances based, user behavior, preferences contributing, energy saving, improved user, user convenience]", - "clean_data": "Alfred Pennyworth\nProduct ManagerSilicon Valley CA USA\n\u2642\u00b6obilealt\n/envel\u2322\n/linkedininapennyworth\n\nProfessional Summary\nSeasoned Product Manager with over 20 years of experience in software development and product\nmanagement having worked at all FAANG companies Exceptional leadership skills strategic\nthinking and a track record of managing products from conception to market success\nSkills\nProduct management Agile methodologies Leadership Communication Project\nmanagement User Experience Design Market Research Data Analysis Java\nPython JavaScript HTML/CSS SQL AWS\nExperience\n2017 \nPresentProduct Manager Google Mountain View CA USA\nLeading crossfunctional teams to design develop and launch innovative products Devel\noping product strategies and making datadriven decisions to improve user experience and\nmeet business goals\n2012 2017 Software Development Engineer III Amazon Seattle WA USA\nLed a team of developers in building scalable and highperforming ecommerce applications\nSuccessfully delivered multiple projects within the stipulated time and budget\n2007 2012 Software Development Engineer II Apple Cupertino CA USA\nDesigned and implemented software components for various Apple services Optimized the\nperformance of applications and improved code quality through thorough testing\n2002 2007 Software Development Engineer I Netflix Los Gatos CA USA\nDeveloped and maintained the user interface for the Netflix web application Worked closely\nwith product managers and designers to create an optimal user experience\n1999 2002 Software Development Engineer I Facebook Menlo Park CA USA\nPlayed a key role in the development of early Facebook features Implemented scalable\nbackend services using Java and SQL\nEducation\n2016 2018 Master of Business Administration Stanford University Stanford CA USA\n1997 1999 Master of Science in Computer Science Massachusetts Institute of Technology \nCambridge MA USA\n1994 1997 Bachelor of Science in Computer Science University of California Berkeley \nBerkeley CA USA\nProjects\n1/2 2020 \nPresentPersonal Project Home Automation System\nDeveloped a smart home automation system using Raspberry Pi and Python The system\nautomates various home appliances based on user behavior and preferences contributing to\nenergy saving and improved user convenience\n2/2", - "emails": [ - "pealfred.pennyworth@email.com" - ], - "entities": [ - "Netflix", - "PresentProduct", - "Devel", - "SQL", - "Seattle WA USA", - "Software Development Engineer I Netflix", - "PresentPersonal Project Home Automation System", - "Los Gatos", - "Business Administration Stanford University", - "Cambridge MA USA", - "Stanford CA USA", - "Amazon", - "Berkeley CA USA", - "Apple", - "Education", - "User Experience Design Market Research Data" - ], - "experience": "Summary \n Seasoned Product Manager with over 20 years of experience in software development and product \n management having worked at all FAANG companies Exceptional leadership skills strategic \n thinking and a track record of managing products from conception to market success \n Skills \n Product management Agile methodologies Leadership Communication Project \n management User Experience Design Market Research Data Analysis Java \n Python JavaScript HTML / CSS SQL AWS \n Experience \n 2017 \n PresentProduct Manager Google Mountain View CA USA \n Leading crossfunctional teams to design develop and launch innovative products Devel \n oping product strategies and making datadriven decisions to improve user experience and \n meet business goals \n 2012 2017 Software Development Engineer III Amazon Seattle WA USA \n Led a team of developers in building scalable and highperforming ecommerce applications \n Successfully delivered multiple projects within the stipulated time and budget \n 2007 2012 Software Development Engineer II Apple Cupertino CA USA \n Designed and implemented software components for various Apple services Optimized the \n performance of applications and improved code quality through thorough testing \n 2002 2007 Software Development Engineer I Netflix Los Gatos CA USA \n Developed and maintained the user interface for the Netflix web application Worked closely \n with product managers and designers to create an optimal user experience \n 1999 2002 Software Development Engineer I Facebook Menlo Park CA USA \n Played a key role in the development of early Facebook features Implemented scalable \n backend services using Java and SQL \n Education \n 2016 2018 Master of Business Administration Stanford University Stanford CA USA \n 1997 1999 Master of Science in Computer Science Massachusetts Institute of Technology \n Cambridge MA USA \n 1994 1997 Bachelor of Science in Computer Science University of California Berkeley \n Berkeley CA USA \n Projects \n 1/2 2020 \n PresentPersonal Project Home Automation System \n Developed a smart home automation system using Raspberry Pi and Python The system \n automates various home appliances based on user behavior and preferences contributing to \n energy saving and improved user convenience \n 2/2", - "extracted_keywords": [ - "Alfred", - "Pennyworth", - "Product", - "ManagerSilicon", - "Valley", - "CA", - "USA", - "\u2642", - "\u00b6obilealt", - "Professional", - "Summary", - "Seasoned", - "Product", - "Manager", - "years", - "experience", - "software", - "development", - "product", - "management", - "FAANG", - "companies", - "leadership", - "thinking", - "track", - "record", - "products", - "conception", - "market", - "success", - "Skills", - "Product", - "management", - "Agile", - "methodologies", - "Leadership", - "Communication", - "Project", - "management", - "User", - "Experience", - "Design", - "Market", - "Research", - "Data", - "Analysis", - "Java", - "Python", - "JavaScript", - "HTML", - "CSS", - "SQL", - "AWS", - "Experience", - "PresentProduct", - "Manager", - "Google", - "Mountain", - "View", - "CA", - "USA", - "teams", - "products", - "Devel", - "product", - "strategies", - "decisions", - "user", - "experience", - "business", - "goals", - "Software", - "Development", - "Engineer", - "III", - "Amazon", - "Seattle", - "WA", - "USA", - "team", - "developers", - "highperforming", - "ecommerce", - "applications", - "projects", - "time", - "budget", - "Software", - "Development", - "Engineer", - "II", - "Apple", - "Cupertino", - "CA", - "USA", - "software", - "components", - "Apple", - "services", - "performance", - "applications", - "code", - "quality", - "testing", - "Software", - "Development", - "Engineer", - "I", - "Netflix", - "Los", - "Gatos", - "CA", - "USA", - "Developed", - "user", - "interface", - "Netflix", - "web", - "application", - "product", - "managers", - "designers", - "user", - "experience", - "Software", - "Development", - "Engineer", - "Facebook", - "Menlo", - "Park", - "CA", - "USA", - "role", - "development", - "Facebook", - "features", - "services", - "Java", - "SQL", - "Education", - "Master", - "Business", - "Administration", - "Stanford", - "University", - "Stanford", - "CA", - "USA", - "Master", - "Science", - "Computer", - "Science", - "Massachusetts", - "Institute", - "Technology", - "Cambridge", - "MA", - "USA", - "Bachelor", - "Science", - "Computer", - "Science", - "University", - "California", - "Berkeley", - "Berkeley", - "CA", - "USA", - "Projects", - "PresentPersonal", - "Project", - "Home", - "Automation", - "System", - "home", - "automation", - "system", - "Raspberry", - "Pi", - "Python", - "system", - "home", - "appliances", - "user", - "behavior", - "preferences", - "energy", - "saving", - "user", - "convenience" - ], - "keyterms": [ - [ - "Software Development Engineer", - 0.2532879474096427 - ], - [ - "USA", - 0.0908144765704578 - ], - [ - "product", - 0.06002611145320178 - ], - [ - "software", - 0.022530132055106372 - ], - [ - "experience", - 0.02143816867647363 - ], - [ - "management", - 0.02074385569284429 - ], - [ - "development", - 0.01766344418212818 - ], - [ - "user experience", - 0.017386580372690356 - ], - [ - "Manager", - 0.011915797594707535 - ], - [ - "Python", - 0.011856462684933523 - ], - [ - "SQL", - 0.011767339129786123 - ], - [ - "Project", - 0.011743798464230066 - ], - [ - "team", - 0.011515715597603877 - ], - [ - "application", - 0.0100341792709379 - ], - [ - "scalable", - 0.00737109583375892 - ], - [ - "Java", - 0.006971290228384053 - ], - [ - "AWS", - 0.006933520895706676 - ], - [ - "Apple", - 0.006866594203670648 - ], - [ - "leadership", - 0.006827491537536502 - ], - [ - "exceptional", - 0.0067843320435271096 - ] - ], - "name": [ - "Alfred Pennyworth" - ], - "phones": [], - "pos_frequencies": { - "ADJ": 18, - "ADP": 27, - "ADV": 1, - "AUX": 1, - "CCONJ": 16, - "DET": 12, - "NOUN": 74, - "NUM": 19, - "PART": 3, - "PRON": 1, - "PROPN": 117, - "SPACE": 56, - "SYM": 1, - "VERB": 31 - }, - "resume_data": "Alfred Pennyworth\nProduct ManagerSilicon Valley, CA, USA\n\u2642\u00b6obile-alt(123) 456-7890\n/envel\u2322pealfred.pennyworth@email.com\n/linkedin-inapennyworth\n/githubapennyworth\nProfessional Summary\nSeasoned Product Manager with over 20 years of experience in software development and product\nmanagement, having worked at all FAANG companies. Exceptional leadership skills, strategic\nthinking, and a track record of managing products from conception to market success.\nSkills\nProduct management, Agile methodologies, Leadership, Communication, Project\nmanagement, User Experience Design, Market Research, Data Analysis, Java,\nPython, JavaScript, HTML/CSS, SQL, AWS\nExperience\n2017 \u2013\nPresentProduct Manager ,Google , Mountain View, CA, USA\nLeading cross-functional teams to design, develop, and launch innovative products. Devel-\noping product strategies and making data-driven decisions to improve user experience and\nmeet business goals.\n2012 \u2013 2017 Software Development Engineer III ,Amazon , Seattle, WA, USA\nLed a team of developers in building scalable and high-performing e-commerce applications.\nSuccessfully delivered multiple projects within the stipulated time and budget.\n2007 \u2013 2012 Software Development Engineer II ,Apple , Cupertino, CA, USA\nDesigned and implemented software components for various Apple services. Optimized the\nperformance of applications and improved code quality through thorough testing.\n2002 \u2013 2007 Software Development Engineer I ,Netflix , Los Gatos, CA, USA\nDeveloped and maintained the user interface for the Netflix web application. Worked closely\nwith product managers and designers to create an optimal user experience.\n1999 \u2013 2002 Software Development Engineer I ,Facebook , Menlo Park, CA, USA\nPlayed a key role in the development of early Facebook features. Implemented scalable\nback-end services using Java and SQL.\nEducation\n2016 \u2013 2018 Master of Business Administration ,Stanford University , Stanford, CA, USA\n1997 \u2013 1999 Master of Science in Computer Science ,Massachusetts Institute of Technology ,\nCambridge, MA, USA\n1994 \u2013 1997 Bachelor of Science in Computer Science ,University of California, Berkeley ,\nBerkeley, CA, USA\nProjects\n1/2 2020 \u2013\nPresentPersonal Project: Home Automation System\nDeveloped a smart home automation system using Raspberry Pi and Python. The system\nautomates various home appliances based on user behavior and preferences, contributing to\nenergy saving and improved user convenience.\n2/2", - "tri_grams": "[Product ManagerSilicon Valley, Valley CA USA, Seasoned Product Manager, years of experience, experience in software, development and product, management having worked, FAANG companies Exceptional, companies Exceptional leadership, Exceptional leadership skills, leadership skills strategic, record of managing, products from conception, conception to market, Product management Agile, management Agile methodologies, Agile methodologies Leadership, methodologies Leadership Communication, Leadership Communication Project, management User Experience, User Experience Design, Experience Design Market, Design Market Research, Market Research Data, Research Data Analysis, Data Analysis Java, Python JavaScript HTML, CSS SQL AWS, PresentProduct Manager Google, View CA USA, Leading crossfunctional teams, teams to design, develop and launch, launch innovative products, innovative products Devel, oping product strategies, strategies and making, making datadriven decisions, decisions to improve, improve user experience, meet business goals, Software Development Engineer, Development Engineer III, Engineer III Amazon, Seattle WA USA, Led a team, team of developers, developers in building, scalable and highperforming, highperforming ecommerce applications, Successfully delivered multiple, delivered multiple projects, time and budget, Software Development Engineer, Development Engineer II, Engineer II Apple, Cupertino CA USA, Designed and implemented, implemented software components, Apple services Optimized, performance of applications, applications and improved, improved code quality, quality through thorough, Software Development Engineer, Engineer I Netflix, Gatos CA USA, Developed and maintained, maintained the user, Netflix web application, web application Worked, application Worked closely, managers and designers, designers to create, create an optimal, optimal user experience, Software Development Engineer, Engineer I Facebook, Park CA USA, Played a key, development of early, early Facebook features, Facebook features Implemented, features Implemented scalable, services using Java, Java and SQL, Master of Business, Business Administration Stanford, Administration Stanford University, Stanford CA USA, Master of Science, Science in Computer, Computer Science Massachusetts, Science Massachusetts Institute, Institute of Technology, Cambridge MA USA, Bachelor of Science, Science in Computer, Computer Science University, University of California, Berkeley CA USA, PresentPersonal Project Home, Project Home Automation, Home Automation System, Developed a smart, smart home automation, home automation system, system using Raspberry, Pi and Python, Python The system, automates various home, home appliances based, based on user, behavior and preferences, saving and improved, improved user convenience]", - "unique_id": "b36fac00-1d60-49ba-a9e4-33477c928e98", - "years": [] -} \ No newline at end of file diff --git a/Data/Processed/Resumes/Resume-barry_allen_fe.pdfbef9d9b1-fa14-4aa0-8111-6033cceacb5a.json b/Data/Processed/Resumes/Resume-barry_allen_fe.pdfbef9d9b1-fa14-4aa0-8111-6033cceacb5a.json deleted file mode 100644 index 4506310f..00000000 --- a/Data/Processed/Resumes/Resume-barry_allen_fe.pdfbef9d9b1-fa14-4aa0-8111-6033cceacb5a.json +++ /dev/null @@ -1,260 +0,0 @@ -{ - "bi_grams": "[Barry Allen, FrontEnd DeveloperGoogle, DeveloperGoogle HQ, HQ Mountain, Mountain View, \u2642\u00b6obilealt, /envel\u2322, challenging role, FrontEnd Developer, modern web, web technologies, create intuitive, engaging user, user interfaces, BTech Computer, Computer Science, Engineering Indian, Indian Institute, New Delhi, Delhi India, Overall GPA, JavaScript ES6, ES6+, + TypeScript, TypeScript HTML5, HTML5 CSS3, CSS3 Python, Python React, React Redux, Redux Angular, Vuejs Nodejs, Nodejs Expressjs, Expressjs D3js, D3js Git, Git Docker, Docker Webpack, Webpack Babel, Babel Google, Google Cloud, Platform Firebase, Firebase RESTful, RESTful APIs, APIs GraphQL, GraphQL Agile, Agile Development, Development TestDriven, Development Responsive, Responsive Design, Design UI, PresentSoftware Engineer, Engineer FrontEnd, FrontEnd Google, Mountain View, Developing intuitive, engaging user, user interfaces, Redux Working, Working closely, UX designers, implement responsive, accessible web, web design, design Participating, agile development, development processes, processes including, including sprint, sprint planning, code reviews, reviews Collaborating, backend developers, integrate RESTful, RESTful APIs, ensure seamless, seamless data, data flow, Personal Expense, Expense Tracker, personal expense, expense tracker, tracker application, React Redux, Firebase Imple, mented user, user authentication, Firebase Auth, data storage, Firestore Utilized, data visualization, provide users, spending patterns]", - "clean_data": "Barry Allen\nFrontEnd DeveloperGoogle HQ Mountain View CA USA\n\u2642\u00b6obilealt\n/envel\u2322\n/linkedininbwayne\n\nObjective\nSeeking a challenging role as a FrontEnd Developer where I can leverage my knowledge of UI/UX\ndesign and modern web technologies to create intuitive and engaging user interfaces\nEducation\n2018 2022 BTech Computer Science and Engineering Indian Institute of Technology\nDelhi New Delhi India\nOverall GPA 95/10\nSkills\nJavaScript ES6+ TypeScript HTML5 CSS3 Python React Redux Angular\nVuejs Nodejs Expressjs D3js Git Docker Webpack Babel Google Cloud\nPlatform Firebase RESTful APIs GraphQL Agile Development TestDriven\nDevelopment Responsive Design UI/UX\nExperience\nJune 2022 \nPresentSoftware Engineer FrontEnd Google Mountain View CA USA\nDeveloping intuitive and engaging user interfaces using React and Redux Working closely\nwith UX designers to implement responsive and accessible web design Participating in\nagile development processes including sprint planning and code reviews Collaborating with\nbackend developers to integrate RESTful APIs and ensure seamless data flow\nProjects\n2022 Personal Expense Tracker\nDeveloped a personal expense tracker application using React Redux and Firebase Imple\nmented user authentication using Firebase Auth and data storage using Firestore Utilized\nD3js for data visualization to provide users with insights into their spending patterns", - "emails": [ - "peallthe.stars@google.com" - ], - "entities": [ - "React", - "Firestore Utilized", - "Delhi", - "React Redux", - "UI", - "Firebase Auth", - "FrontEnd", - "BTech Computer Science and Engineering Indian Institute of Technology", - "FrontEnd Developer", - "New Delhi India", - "Skills", - "Platform Firebase", - "Agile Development TestDriven\nDevelopment Responsive Design UI/UX", - "PresentSoftware Engineer FrontEnd Google Mountain View CA USA", - "Projects", - "Redux Working", - "Education" - ], - "experience": "Objective \n Seeking a challenging role as a FrontEnd Developer where I can leverage my knowledge of UI / UX \n design and modern web technologies to create intuitive and engaging user interfaces \n Education \n 2018 2022 BTech Computer Science and Engineering Indian Institute of Technology \n Delhi New Delhi India \n Overall GPA 95/10 \n Skills \n JavaScript ES6 + TypeScript HTML5 CSS3 Python React Redux Angular \n Vuejs Nodejs Expressjs D3js Git Docker Webpack Babel Google Cloud \n Platform Firebase RESTful APIs GraphQL Agile Development TestDriven \n Development Responsive Design UI / UX \n Experience \n June 2022 \n PresentSoftware Engineer FrontEnd Google Mountain View CA USA \n Developing intuitive and engaging user interfaces using React and Redux Working closely \n with UX designers to implement responsive and accessible web design Participating in \n agile development processes including sprint planning and code reviews Collaborating with \n backend developers to integrate RESTful APIs and ensure seamless data flow \n Projects \n 2022 Personal Expense Tracker \n Developed a personal expense tracker application using React Redux and Firebase Imple \n mented user authentication using Firebase Auth and data storage using Firestore Utilized \n D3js for data visualization to provide users with insights into their spending patterns", - "extracted_keywords": [ - "Barry", - "Allen", - "FrontEnd", - "DeveloperGoogle", - "HQ", - "Mountain", - "View", - "CA", - "USA", - "\u2642", - "\u00b6obilealt", - "/linkedininbwayne", - "Objective", - "role", - "FrontEnd", - "Developer", - "knowledge", - "UI", - "UX", - "design", - "web", - "technologies", - "user", - "interfaces", - "Education", - "BTech", - "Computer", - "Science", - "Engineering", - "Indian", - "Institute", - "Technology", - "Delhi", - "New", - "Delhi", - "India", - "Overall", - "GPA", - "Skills", - "JavaScript", - "ES6", - "+", - "TypeScript", - "HTML5", - "CSS3", - "Python", - "React", - "Redux", - "Angular", - "Vuejs", - "Nodejs", - "Expressjs", - "Git", - "Docker", - "Webpack", - "Babel", - "Google", - "Cloud", - "Platform", - "Firebase", - "APIs", - "GraphQL", - "Agile", - "Development", - "TestDriven", - "Development", - "Responsive", - "Design", - "UI", - "UX", - "Experience", - "June", - "PresentSoftware", - "Engineer", - "FrontEnd", - "Google", - "Mountain", - "View", - "CA", - "USA", - "user", - "interfaces", - "React", - "Redux", - "Working", - "UX", - "designers", - "web", - "design", - "development", - "processes", - "sprint", - "planning", - "code", - "reviews", - "developers", - "APIs", - "data", - "flow", - "Projects", - "Personal", - "Expense", - "Tracker", - "expense", - "tracker", - "application", - "React", - "Redux", - "Firebase", - "Imple", - "user", - "authentication", - "Firebase", - "Auth", - "data", - "storage", - "Firestore", - "D3js", - "data", - "visualization", - "users", - "insights", - "spending", - "patterns" - ], - "keyterms": [ - [ - "Mountain View CA USA", - 0.508138095252526 - ], - [ - "FrontEnd", - 0.04217636079129916 - ], - [ - "user interface", - 0.030073321065429096 - ], - [ - "React Redux", - 0.01732513490526635 - ], - [ - "developer", - 0.01694538047555922 - ], - [ - "UX", - 0.016894308063632554 - ], - [ - "restful api", - 0.01576075166735997 - ], - [ - "UI", - 0.014513278517489854 - ], - [ - "web", - 0.012447700730893765 - ], - [ - "intuitive", - 0.011952455914068402 - ], - [ - "Firebase", - 0.008593408253817738 - ], - [ - "D3js", - 0.007871502682311644 - ], - [ - "design", - 0.0073138382759099395 - ], - [ - "Delhi", - 0.0062576095975739035 - ], - [ - "Development", - 0.005933160755540573 - ], - [ - "technology", - 0.004578358830990269 - ], - [ - "DeveloperGoogle", - 0.004561509883154723 - ], - [ - "HQ", - 0.004416248521150768 - ], - [ - "education", - 0.004404513172054677 - ], - [ - "modern", - 0.00422178063142237 - ] - ], - "name": [ - "Barry Allen" - ], - "phones": [], - "pos_frequencies": { - "ADJ": 11, - "ADP": 8, - "ADV": 1, - "AUX": 1, - "CCONJ": 11, - "DET": 3, - "NOUN": 47, - "NUM": 5, - "PART": 4, - "PRON": 3, - "PROPN": 82, - "SCONJ": 2, - "SPACE": 31, - "SYM": 2, - "VERB": 18 - }, - "resume_data": "Barry Allen\nFront-End DeveloperGoogle HQ, Mountain View, CA, USA\n\u2642\u00b6obile-alt(123) 456-7890\n/envel\u2322peallthe.stars@google.com\n/linkedin-inbwayne\n/githubbwayne\nObjective\nSeeking a challenging role as a Front-End Developer where I can leverage my knowledge of UI/UX\ndesign and modern web technologies to create intuitive and engaging user interfaces.\nEducation\n2018 \u2013 2022 B.Tech, Computer Science and Engineering ,Indian Institute of Technology\nDelhi , New Delhi, India\nOverall GPA: 9.5/10\nSkills\nJavaScript (ES6+), TypeScript, HTML5, CSS3, Python, React, Redux, Angular,\nVue.js, Node.js, Express.js, D3.js, Git, Docker, Webpack, Babel, Google Cloud\nPlatform, Firebase, RESTful APIs, GraphQL, Agile Development, Test-Driven\nDevelopment, Responsive Design, UI/UX\nExperience\nJune 2022 \u2013\nPresentSoftware Engineer (Front-End) ,Google , Mountain View, CA, USA\nDeveloping intuitive and engaging user interfaces using React and Redux. Working closely\nwith UX designers to implement responsive and accessible web design. Participating in\nagile development processes, including sprint planning and code reviews. Collaborating with\nback-end developers to integrate RESTful APIs and ensure seamless data flow.\nProjects\n2022 Personal Expense Tracker\nDeveloped a personal expense tracker application using React, Redux, and Firebase. Imple-\nmented user authentication using Firebase Auth and data storage using Firestore. Utilized\nD3.js for data visualization to provide users with insights into their spending patterns.", - "tri_grams": "[FrontEnd DeveloperGoogle HQ, DeveloperGoogle HQ Mountain, HQ Mountain View, View CA USA, Seeking a challenging, leverage my knowledge, knowledge of UI, design and modern, modern web technologies, technologies to create, intuitive and engaging, engaging user interfaces, BTech Computer Science, Science and Engineering, Engineering Indian Institute, Institute of Technology, New Delhi India, JavaScript ES6+, ES6+ TypeScript, + TypeScript HTML5, TypeScript HTML5 CSS3, HTML5 CSS3 Python, CSS3 Python React, Python React Redux, React Redux Angular, Vuejs Nodejs Expressjs, Nodejs Expressjs D3js, Expressjs D3js Git, D3js Git Docker, Git Docker Webpack, Docker Webpack Babel, Webpack Babel Google, Babel Google Cloud, Platform Firebase RESTful, Firebase RESTful APIs, RESTful APIs GraphQL, APIs GraphQL Agile, GraphQL Agile Development, Agile Development TestDriven, Development Responsive Design, Responsive Design UI, PresentSoftware Engineer FrontEnd, Engineer FrontEnd Google, View CA USA, intuitive and engaging, engaging user interfaces, interfaces using React, React and Redux, Redux Working closely, designers to implement, responsive and accessible, accessible web design, web design Participating, agile development processes, development processes including, processes including sprint, including sprint planning, planning and code, code reviews Collaborating, developers to integrate, integrate RESTful APIs, APIs and ensure, ensure seamless data, seamless data flow, Personal Expense Tracker, Developed a personal, personal expense tracker, expense tracker application, application using React, Redux and Firebase, mented user authentication, authentication using Firebase, Auth and data, storage using Firestore, D3js for data, visualization to provide, users with insights]", - "unique_id": "bef9d9b1-fa14-4aa0-8111-6033cceacb5a", - "years": [] -} \ No newline at end of file diff --git a/Data/Processed/Resumes/Resume-bruce_wayne_fullstack.pdf04c228e3-04a4-4970-a149-a3cffc92aee8.json b/Data/Processed/Resumes/Resume-bruce_wayne_fullstack.pdf04c228e3-04a4-4970-a149-a3cffc92aee8.json deleted file mode 100644 index 328ca6dc..00000000 --- a/Data/Processed/Resumes/Resume-bruce_wayne_fullstack.pdf04c228e3-04a4-4970-a149-a3cffc92aee8.json +++ /dev/null @@ -1,345 +0,0 @@ -{ - "bi_grams": "[Bruce Wayne, MERN Stack, Stack Developer123, Developer123 Gotham, Gotham St, Gotham USA, \u2642\u00b6obilealt, /envel\u2322, Professional Summary, Highly skilled, skilled MERN, MERN Stack, Stack Developer, experience specializing, maintaining complex, complex web, web applications, applications Proficient, MongoDB Expressjs, Expressjs React, Nodejs Currently, Currently contributing, AI technologies, ChatGPT project, TypeScript MongoDB, MongoDB Expressjs, Expressjs React, React Nodejs, Nodejs MERN, MERN stack, RESTful APIs, APIs Git, GitHub Docker, Kubernetes Agile, Scrum Python, Machine Learning, Learning basics, PresentMERN Stack, Stack Developer, Developer OpenAI, San Francisco, Francisco USA, ChatGPT project, Nodejs Expressjs, Implementing RESTful, RESTful services, backend Utilizing, applications Working, Agile environment, environment delivering, delivering highquality, highquality software, sprint Contributing, machine learning, learning algorithms, natural language, language processing, processing tasks, 2020Full Stack, Stack Developer, Developer Uber, San Francisco, Francisco USA, maintained scalable, scalable web, web applications, MERN stack, stack Ensured, performance quality, applications Successfully, Successfully deployed, deployed solutions, Kubernetes Collaborated, engineers product, product managers, UX designers, designers Led, junior developers, developers conducted, conducted code, code reviews, best coding, coding practices, practices Worked, Worked closely, data science, science team, recommendation algorithms, enhance user, user experience, June 2015Software, 2015Software Developer, Developer Facebook, Menlo Park, Park USA, Developed features, Facebook web, web application, React Ensured, MongoDB databases, databases Utilized, Utilized RESTful, RESTful APIs, application Worked, fastpaced testdriven, testdriven development, development environment, legacy system, modern MERN, MERN stack, stack architecture, Computer Science, Science CalTech, Pasadena USA, Computer Science, Science MIT, MIT Cambridge, Cambridge USA, Computer Science, Science UC, UC San, San Diego, San Diego, Diego USA, PresentPersonal Project, Project Gotham, Gotham Event, Event Planner, fullfeatured web, web application, organize events, Gotham city, MERN stack, application allows, allows users, create manage, share events, Google Maps, Maps API, display event]", - "clean_data": "Bruce Wayne\nMERN Stack Developer123 Gotham St\nGotham USA\n\u2642\u00b6obilealt\n/envel\u2322\n/linkedininbwayne\n\nProfessional Summary\nHighly skilled MERN Stack Developer with over 10 years of experience specializing in designing\nbuilding and maintaining complex web applications Proficient in MongoDB Expressjs React and\nNodejs Currently contributing to the development of AI technologies at OpenAI with a primary\nfocus on the ChatGPT project\nSkills\nJavaScript and TypeScript MongoDB Expressjs React Nodejs MERN stack\nRESTful APIs Git and GitHub Docker and Kubernetes Agile and Scrum Python\nand Machine Learning basics\nExperience\nJune 2020 \nPresentMERN Stack Developer OpenAI San Francisco USA\nWorking on the development of the ChatGPT project using Nodejs Expressjs and React\nImplementing RESTful services for communication between frontend and backend Utilizing\nDocker and Kubernetes for deployment and management of applications Working in an\nAgile environment delivering highquality software every sprint Contributing to the design\nand implementation of machine learning algorithms for natural language processing tasks\nJuly 2015 \nMay 2020Full Stack Developer Uber San Francisco USA\nDeveloped and maintained scalable web applications using MERN stack Ensured the\nperformance quality and responsiveness of applications Successfully deployed solutions\nusing Docker and Kubernetes Collaborated with a team of engineers product managers\nand UX designers Led a team of junior developers conducted code reviews and ensured\nadherence to best coding practices Worked closely with the data science team to optimize\nrecommendation algorithms and enhance user experience\nJune 2012 \nJune 2015Software Developer Facebook Menlo Park USA\nDeveloped features for the Facebook web application using React Ensured the performance\nof the MongoDB databases Utilized RESTful APIs for communication between different\nparts of the application Worked in a fastpaced testdriven development environment\nAssisted in migrating the legacy system to a modern MERN stack architecture\nEducation\n2009 2012 PhD in Computer Science CalTech Pasadena USA\n2007 2009 Master of Science in Computer Science MIT Cambridge USA\n2003 2007 Bachelor of Science in Computer Science UC San Diego San Diego USA\n1/2 Projects\n2019 \nPresentPersonal Project Gotham Event Planner\nCreated a fullfeatured web application to plan and organize events in Gotham city Used\nMERN stack for development and Docker for deployment The application allows users to\ncreate manage and share events and integrates with Google Maps API to display event\nlocations\n2/2", - "emails": [ - "pebwayne@email.com" - ], - "entities": [ - "Kubernetes Collaborated", - "Working", - "Bachelor of Science in Computer Science UC", - "Assisted", - "AI", - "Utilized", - "Kubernetes", - "GitHub Docker", - "OpenAI", - "CalTech", - "React Ensured", - "Google Maps API", - "Education", - "UX", - "Nodejs", - "PresentPersonal", - "TypeScript", - "Git", - "Utilizing", - "Skills", - "Machine Learning", - "Nodejs Expressjs and React\nImplementing RESTful services", - "Pasadena USA" - ], - "experience": "Summary \n Highly skilled MERN Stack Developer with over 10 years of experience specializing in designing \n building and maintaining complex web applications Proficient in MongoDB Expressjs React and \n Nodejs Currently contributing to the development of AI technologies at OpenAI with a primary \n focus on the ChatGPT project \n Skills \n JavaScript and TypeScript MongoDB Expressjs React Nodejs MERN stack \n RESTful APIs Git and GitHub Docker and Kubernetes Agile and Scrum Python \n and Machine Learning basics \n Experience \n June 2020 \n PresentMERN Stack Developer OpenAI San Francisco USA \n Working on the development of the ChatGPT project using Nodejs Expressjs and React \n Implementing RESTful services for communication between frontend and backend Utilizing \n Docker and Kubernetes for deployment and management of applications Working in an \n Agile environment delivering highquality software every sprint Contributing to the design \n and implementation of machine learning algorithms for natural language processing tasks \n July 2015 \n May 2020Full Stack Developer Uber San Francisco USA \n Developed and maintained scalable web applications using MERN stack Ensured the \n performance quality and responsiveness of applications Successfully deployed solutions \n using Docker and Kubernetes Collaborated with a team of engineers product managers \n and UX designers Led a team of junior developers conducted code reviews and ensured \n adherence to best coding practices Worked closely with the data science team to optimize \n recommendation algorithms and enhance user experience \n June 2012 \n June 2015Software Developer Facebook Menlo Park USA \n Developed features for the Facebook web application using React Ensured the performance \n of the MongoDB databases Utilized RESTful APIs for communication between different \n parts of the application Worked in a fastpaced testdriven development environment \n Assisted in migrating the legacy system to a modern MERN stack architecture \n Education \n 2009 2012 PhD in Computer Science CalTech Pasadena USA \n 2007 2009 Master of Science in Computer Science MIT Cambridge USA \n 2003 2007 Bachelor of Science in Computer Science UC San Diego San Diego USA \n 1/2 Projects \n 2019 \n PresentPersonal Project Gotham Event Planner \n Created a fullfeatured web application to plan and organize events in Gotham city Used \n MERN stack for development and Docker for deployment The application allows users to \n create manage and share events and integrates with Google Maps API to display event \n locations \n 2/2", - "extracted_keywords": [ - "Bruce", - "Wayne", - "MERN", - "Stack", - "Developer123", - "Gotham", - "St", - "Gotham", - "USA", - "\u2642", - "\u00b6obilealt", - "/linkedininbwayne", - "Professional", - "Summary", - "MERN", - "Stack", - "Developer", - "years", - "experience", - "building", - "web", - "applications", - "Expressjs", - "React", - "Nodejs", - "development", - "AI", - "technologies", - "OpenAI", - "focus", - "ChatGPT", - "project", - "Skills", - "JavaScript", - "TypeScript", - "MongoDB", - "Expressjs", - "React", - "Nodejs", - "MERN", - "APIs", - "Git", - "GitHub", - "Docker", - "Kubernetes", - "Agile", - "Scrum", - "Python", - "Machine", - "Learning", - "basics", - "Experience", - "June", - "PresentMERN", - "Stack", - "Developer", - "OpenAI", - "San", - "Francisco", - "USA", - "Working", - "development", - "ChatGPT", - "project", - "Nodejs", - "Expressjs", - "React", - "services", - "communication", - "frontend", - "Utilizing", - "Docker", - "Kubernetes", - "deployment", - "management", - "applications", - "environment", - "highquality", - "software", - "sprint", - "design", - "implementation", - "machine", - "algorithms", - "language", - "processing", - "tasks", - "July", - "May", - "2020Full", - "Stack", - "Developer", - "Uber", - "San", - "Francisco", - "USA", - "Developed", - "web", - "applications", - "MERN", - "stack", - "performance", - "quality", - "responsiveness", - "applications", - "solutions", - "Docker", - "Kubernetes", - "Collaborated", - "team", - "engineers", - "product", - "managers", - "UX", - "designers", - "team", - "developers", - "code", - "reviews", - "adherence", - "practices", - "data", - "science", - "team", - "recommendation", - "algorithms", - "user", - "experience", - "June", - "June", - "Developer", - "Facebook", - "Menlo", - "Park", - "USA", - "features", - "Facebook", - "web", - "application", - "React", - "performance", - "MongoDB", - "APIs", - "communication", - "parts", - "application", - "testdriven", - "development", - "environment", - "legacy", - "system", - "MERN", - "stack", - "architecture", - "Education", - "PhD", - "Computer", - "Science", - "CalTech", - "Pasadena", - "USA", - "Master", - "Science", - "Computer", - "Science", - "MIT", - "Cambridge", - "USA", - "Bachelor", - "Science", - "Computer", - "Science", - "UC", - "San", - "Diego", - "San", - "Diego", - "USA", - "Projects", - "PresentPersonal", - "Project", - "Gotham", - "Event", - "Planner", - "web", - "application", - "events", - "Gotham", - "city", - "MERN", - "stack", - "development", - "Docker", - "deployment", - "application", - "users", - "manage", - "events", - "integrates", - "Google", - "Maps", - "API", - "event", - "locations" - ], - "keyterms": [ - [ - "San Francisco USA", - 0.28104777254672486 - ], - [ - "web application", - 0.07395838902227116 - ], - [ - "Expressjs React", - 0.06672851125833305 - ], - [ - "chatgpt project", - 0.06476011153803286 - ], - [ - "MERN stack", - 0.05362165690837463 - ], - [ - "restful api", - 0.04329931776733209 - ], - [ - "developer", - 0.016905339657325078 - ], - [ - "Gotham", - 0.015639185899381697 - ], - [ - "experience", - 0.01451706237646984 - ], - [ - "Nodejs", - 0.011001663128927773 - ], - [ - "development", - 0.010920186599081536 - ], - [ - "Docker", - 0.010051444050915393 - ], - [ - "Kubernetes", - 0.009983141026320433 - ], - [ - "Stack", - 0.008781417492375376 - ], - [ - "OpenAI", - 0.007807257743537287 - ], - [ - "June", - 0.0073390704491771526 - ], - [ - "Computer Science", - 0.006560451122224374 - ], - [ - "communication", - 0.005158863753107799 - ], - [ - "environment", - 0.004602922915578097 - ], - [ - "algorithm", - 0.004432429485404663 - ] - ], - "name": [ - "Bruce Wayne" - ], - "phones": [], - "pos_frequencies": { - "ADJ": 18, - "ADP": 42, - "ADV": 3, - "AUX": 1, - "CCONJ": 22, - "DET": 21, - "NOUN": 99, - "NUM": 18, - "PART": 4, - "PROPN": 104, - "SCONJ": 1, - "SPACE": 56, - "SYM": 1, - "VERB": 37 - }, - "resume_data": "Bruce Wayne\nMERN Stack Developer123 Gotham St\nGotham, USA\n\u2642\u00b6obile-alt(123) 456-7890\n/envel\u2322pebwayne@email.com\n/linkedin-inbwayne\n/githubbwayne\nProfessional Summary\nHighly skilled MERN Stack Developer with over 10 years of experience, specializing in designing,\nbuilding, and maintaining complex web applications. Proficient in MongoDB, Express.js, React, and\nNode.js. Currently contributing to the development of AI technologies at OpenAI, with a primary\nfocus on the ChatGPT project.\nSkills\nJavaScript and TypeScript, MongoDB, Express.js, React, Node.js (MERN stack),\nRESTful APIs, Git and GitHub, Docker and Kubernetes, Agile and Scrum, Python\nand Machine Learning basics\nExperience\nJune 2020 \u2013\nPresentMERN Stack Developer ,OpenAI , San Francisco, USA\nWorking on the development of the ChatGPT project, using Node.js, Express.js, and React.\nImplementing RESTful services for communication between frontend and backend. Utilizing\nDocker and Kubernetes for deployment and management of applications. Working in an\nAgile environment, delivering high-quality software every sprint. Contributing to the design\nand implementation of machine learning algorithms for natural language processing tasks.\nJuly 2015 \u2013\nMay 2020Full Stack Developer ,Uber , San Francisco, USA\nDeveloped and maintained scalable web applications using MERN stack. Ensured the\nperformance, quality, and responsiveness of applications. Successfully deployed solutions\nusing Docker and Kubernetes. Collaborated with a team of engineers, product managers,\nand UX designers. Led a team of junior developers, conducted code reviews, and ensured\nadherence to best coding practices. Worked closely with the data science team to optimize\nrecommendation algorithms and enhance user experience.\nJune 2012 \u2013\nJune 2015Software Developer ,Facebook , Menlo Park, USA\nDeveloped features for the Facebook web application using React. Ensured the performance\nof the MongoDB databases. Utilized RESTful APIs for communication between different\nparts of the application. Worked in a fast-paced, test-driven development environment.\nAssisted in migrating the legacy system to a modern MERN stack architecture.\nEducation\n2009 \u2013 2012 PhD in Computer Science ,CalTech , Pasadena, USA\n2007 \u2013 2009 Master of Science in Computer Science ,MIT, Cambridge, USA\n2003 \u2013 2007 Bachelor of Science in Computer Science ,UC San Diego , San Diego, USA\n1/2 Projects\n2019 \u2013\nPresentPersonal Project: Gotham Event Planner\nCreated a full-featured web application to plan and organize events in Gotham city. Used\nMERN stack for development and Docker for deployment. The application allows users to\ncreate, manage, and share events, and integrates with Google Maps API to display event\nlocations.\n2/2", - "tri_grams": "[MERN Stack Developer123, Stack Developer123 Gotham, Developer123 Gotham St, Highly skilled MERN, skilled MERN Stack, MERN Stack Developer, years of experience, specializing in designing, building and maintaining, maintaining complex web, complex web applications, web applications Proficient, Proficient in MongoDB, MongoDB Expressjs React, Nodejs Currently contributing, development of AI, technologies at OpenAI, JavaScript and TypeScript, TypeScript MongoDB Expressjs, MongoDB Expressjs React, Expressjs React Nodejs, React Nodejs MERN, Nodejs MERN stack, RESTful APIs Git, Git and GitHub, Docker and Kubernetes, Agile and Scrum, Machine Learning basics, PresentMERN Stack Developer, Stack Developer OpenAI, San Francisco USA, project using Nodejs, Expressjs and React, Implementing RESTful services, services for communication, communication between frontend, frontend and backend, Docker and Kubernetes, Kubernetes for deployment, deployment and management, management of applications, Agile environment delivering, environment delivering highquality, delivering highquality software, software every sprint, implementation of machine, machine learning algorithms, algorithms for natural, natural language processing, language processing tasks, 2020Full Stack Developer, Stack Developer Uber, San Francisco USA, Developed and maintained, maintained scalable web, scalable web applications, applications using MERN, MERN stack Ensured, quality and responsiveness, responsiveness of applications, applications Successfully deployed, Successfully deployed solutions, Docker and Kubernetes, team of engineers, engineers product managers, UX designers Led, Led a team, team of junior, junior developers conducted, developers conducted code, conducted code reviews, reviews and ensured, adherence to best, best coding practices, coding practices Worked, practices Worked closely, data science team, team to optimize, algorithms and enhance, enhance user experience, June 2015Software Developer, 2015Software Developer Facebook, Menlo Park USA, Facebook web application, application using React, Ensured the performance, MongoDB databases Utilized, databases Utilized RESTful, Utilized RESTful APIs, APIs for communication, communication between different, fastpaced testdriven development, testdriven development environment, Assisted in migrating, migrating the legacy, modern MERN stack, MERN stack architecture, PhD in Computer, Computer Science CalTech, Master of Science, Science in Computer, Computer Science MIT, Science MIT Cambridge, MIT Cambridge USA, Bachelor of Science, Science in Computer, Computer Science UC, Science UC San, UC San Diego, San Diego USA, PresentPersonal Project Gotham, Project Gotham Event, Gotham Event Planner, Created a fullfeatured, fullfeatured web application, application to plan, plan and organize, events in Gotham, stack for development, development and Docker, Docker for deployment, deployment The application, application allows users, manage and share, events and integrates, integrates with Google, Google Maps API, API to display]", - "unique_id": "04c228e3-04a4-4970-a149-a3cffc92aee8", - "years": [] -} \ No newline at end of file diff --git a/Data/Processed/Resumes/Resume-harvey_dent_mle.pdf5dde9183-5af4-4107-91e1-7b2d66fec490.json b/Data/Processed/Resumes/Resume-harvey_dent_mle.pdf5dde9183-5af4-4107-91e1-7b2d66fec490.json deleted file mode 100644 index a10663c1..00000000 --- a/Data/Processed/Resumes/Resume-harvey_dent_mle.pdf5dde9183-5af4-4107-91e1-7b2d66fec490.json +++ /dev/null @@ -1,304 +0,0 @@ -{ - "bi_grams": "[Harvey Dent, Machine Learning, Learning Engineer321, Engineer321 Gotham, Gotham St, Gotham USA, \u2642\u00b6obilealt, /envel\u2322, Professional Summary, Machine Learning, Learning Engineer, designing building, deploying predictive, deep learning, learning systems, systems Proficient, Python TensorFlow, TensorFlow PyTorch, Currently leading, AI engineers, Python R, R TensorFlow, TensorFlow PyTorch, PyTorch Scikitlearn, Scikitlearn Keras, Keras SQL, SQL NoSQL, NoSQL Git, Git Docker, Kubernetes Agile, Scrum Statistics, Statistics Data, Data visualization, visualization Deep, Deep Learning, Learning Natural, Language Processing, PresentMachine Learning, Learning Engineer, Engineer OpenAI, San Francisco, Francisco USA, AI engineers, engineers Designed, implemented deep, deep learning, learning models, language processing, processing tasks, tasks Improved, model training, data processing, pipelines Published, research papers, toptier AI, AI conferences, Data Scientist, Scientist Uber, San Francisco, Francisco USA, deployed machine, machine learning, learning models, ride allocation, algorithms Utilized, Utilized TensorFlow, developing predictive, predictive models, models Analyzed, visualized large, large data, data sets, drive business, business strategies, Junior Data, Data Analyst, Analyst Facebook, Menlo Park, Park USA, visualized large, large datasets, R Assisted, machine learning, learning models, user behavior, behavior prediction, prediction Conducted, B testing, product team, Computer Science, Science Specialization, AI MIT, MIT Cambridge, Computer Science, Science UC, UC San, San Diego, San Diego, Diego USA, PresentPersonal Project, Project Predictive, Predictive Stock, Stock Trading, Trading Model, predictive stock, stock trading, trading model, deep learning, time series, series analysis, model development, model predicts, predicts stock, high degree, automates trading, trading decisions]", - "clean_data": "Harvey Dent\nMachine Learning Engineer321 Gotham St\nGotham USA\n\u2642\u00b6obilealt\n/envel\u2322\n/linkedininhdent\n/githubhdent\nProfessional Summary\nMachine Learning Engineer with 5 years of experience in designing building and deploying predictive\nmodels and deep learning systems Proficient in Python TensorFlow PyTorch and Scikitlearn\nCurrently leading a team of AI engineers at OpenAI\nSkills\nPython R TensorFlow PyTorch Scikitlearn Keras SQL NoSQL Git Docker\nKubernetes Agile and Scrum Statistics Data visualization Deep Learning Natural\nLanguage Processing\nExperience\n2021 \nPresentMachine Learning Engineer OpenAI San Francisco USA\nLeading a team of AI engineers Designed and implemented deep learning models for natural\nlanguage processing tasks Improved the efficiency of model training and data processing\npipelines Published several research papers in toptier AI conferences\n2018 2021 Data Scientist Uber San Francisco USA\nDeveloped and deployed machine learning models to improve the efficiency of ride allocation\nalgorithms Utilized TensorFlow and PyTorch for developing predictive models Analyzed\nand visualized large data sets to drive business strategies\n2016 2018 Junior Data Analyst Facebook Menlo Park USA\nAnalyzed and visualized large datasets using Python and R Assisted in the development of\nmachine learning models for user behavior prediction Conducted A/B testing and provided\ninsights to the product team\nEducation\n2014 2016 Master of Science in Computer Science Specialization in AI MIT Cambridge\nUSA\n2010 2014 Bachelor of Science in Computer Science UC San Diego San Diego USA\nProjects\n2021 \nPresentPersonal Project Predictive Stock Trading Model\nDeveloped a predictive stock trading model using deep learning and time series analysis\nUsed PyTorch for model development and Docker for deployment The model predicts stock\nprices with a high degree of accuracy and automates trading decisions", - "emails": [ - "pehdent@email.com" - ], - "entities": [ - "Education", - "SQL", - "Conducted A/B", - "Cambridge", - "PresentPersonal Project Predictive Stock Trading Model", - "PresentMachine Learning Engineer OpenAI", - "AI", - "Utilized", - "Harvey Dent\nMachine Learning", - "San Francisco USA", - "Science in Computer Science Specialization", - "USA", - "Projects", - "Menlo Park USA", - "Python TensorFlow PyTorch", - "Statistics Data", - "Data Scientist Uber San Francisco USA" - ], - "experience": "Summary \n Machine Learning Engineer with 5 years of experience in designing building and deploying predictive \n models and deep learning systems Proficient in Python TensorFlow PyTorch and Scikitlearn \n Currently leading a team of AI engineers at OpenAI \n Skills \n Python R TensorFlow PyTorch Scikitlearn Keras SQL NoSQL Git Docker \n Kubernetes Agile and Scrum Statistics Data visualization Deep Learning Natural \n Language Processing \n Experience \n 2021 \n PresentMachine Learning Engineer OpenAI San Francisco USA \n Leading a team of AI engineers Designed and implemented deep learning models for natural \n language processing tasks Improved the efficiency of model training and data processing \n pipelines Published several research papers in toptier AI conferences \n 2018 2021 Data Scientist Uber San Francisco USA \n Developed and deployed machine learning models to improve the efficiency of ride allocation \n algorithms Utilized TensorFlow and PyTorch for developing predictive models Analyzed \n and visualized large data sets to drive business strategies \n 2016 2018 Junior Data Analyst Facebook Menlo Park USA \n Analyzed and visualized large datasets using Python and R Assisted in the development of \n machine learning models for user behavior prediction Conducted A / B testing and provided \n insights to the product team \n Education \n 2014 2016 Master of Science in Computer Science Specialization in AI MIT Cambridge \n USA \n 2010 2014 Bachelor of Science in Computer Science UC San Diego San Diego USA \n Projects \n 2021 \n PresentPersonal Project Predictive Stock Trading Model \n Developed a predictive stock trading model using deep learning and time series analysis \n Used PyTorch for model development and Docker for deployment The model predicts stock \n prices with a high degree of accuracy and automates trading decisions", - "extracted_keywords": [ - "Harvey", - "Dent", - "Machine", - "Learning", - "Engineer321", - "Gotham", - "St", - "Gotham", - "USA", - "\u2642", - "\u00b6obilealt", - "/linkedininhdent", - "/githubhdent", - "Professional", - "Summary", - "Machine", - "Learning", - "Engineer", - "years", - "experience", - "building", - "models", - "learning", - "systems", - "Proficient", - "Python", - "TensorFlow", - "PyTorch", - "Scikitlearn", - "team", - "AI", - "engineers", - "OpenAI", - "Skills", - "Python", - "R", - "TensorFlow", - "PyTorch", - "Scikitlearn", - "Keras", - "SQL", - "NoSQL", - "Git", - "Docker", - "Kubernetes", - "Agile", - "Scrum", - "Statistics", - "Data", - "visualization", - "Deep", - "Learning", - "Natural", - "Language", - "Processing", - "Experience", - "PresentMachine", - "Learning", - "Engineer", - "OpenAI", - "San", - "Francisco", - "USA", - "team", - "AI", - "engineers", - "learning", - "models", - "language", - "processing", - "tasks", - "efficiency", - "model", - "training", - "data", - "processing", - "pipelines", - "research", - "papers", - "toptier", - "AI", - "conferences", - "Data", - "Scientist", - "Uber", - "San", - "Francisco", - "USA", - "Developed", - "machine", - "learning", - "models", - "efficiency", - "ride", - "allocation", - "TensorFlow", - "PyTorch", - "models", - "data", - "sets", - "business", - "strategies", - "Junior", - "Data", - "Analyst", - "Facebook", - "Menlo", - "Park", - "USA", - "datasets", - "Python", - "R", - "Assisted", - "development", - "machine", - "learning", - "models", - "user", - "behavior", - "prediction", - "Conducted", - "A", - "B", - "testing", - "insights", - "product", - "team", - "Education", - "Master", - "Science", - "Computer", - "Science", - "Specialization", - "AI", - "MIT", - "Cambridge", - "USA", - "Bachelor", - "Science", - "Computer", - "Science", - "UC", - "San", - "Diego", - "San", - "Diego", - "USA", - "Projects", - "PresentPersonal", - "Project", - "Predictive", - "Stock", - "Trading", - "Model", - "stock", - "trading", - "model", - "learning", - "time", - "series", - "analysis", - "PyTorch", - "model", - "development", - "Docker", - "deployment", - "model", - "stock", - "prices", - "degree", - "accuracy", - "trading", - "decisions" - ], - "keyterms": [ - [ - "machine learning model", - 0.29384450152934605 - ], - [ - "San Francisco USA", - 0.08868011510384628 - ], - [ - "Learning Engineer", - 0.07071805326147397 - ], - [ - "TensorFlow PyTorch", - 0.0664518697682042 - ], - [ - "deep learning", - 0.0613735746540308 - ], - [ - "AI engineer", - 0.04837101835268983 - ], - [ - "Machine Learning", - 0.03209928475884996 - ], - [ - "Python", - 0.01406659130079881 - ], - [ - "team", - 0.011712162383831456 - ], - [ - "predictive", - 0.011287884257850087 - ], - [ - "experience", - 0.010376334392954044 - ], - [ - "Data", - 0.009787022279697168 - ], - [ - "processing", - 0.0083226092101047 - ], - [ - "OpenAI", - 0.007625692166684825 - ], - [ - "Docker", - 0.006700131357292339 - ], - [ - "Scikitlearn", - 0.006458217841891463 - ], - [ - "Computer Science", - 0.005378059431970814 - ], - [ - "datum", - 0.005076972017622222 - ], - [ - "efficiency", - 0.0050045146499247805 - ], - [ - "San Diego", - 0.0044666168358042145 - ] - ], - "name": [], - "phones": [], - "pos_frequencies": { - "ADJ": 13, - "ADP": 25, - "ADV": 1, - "CCONJ": 15, - "DET": 9, - "NOUN": 73, - "NUM": 11, - "PART": 2, - "PROPN": 99, - "SPACE": 46, - "SYM": 2, - "VERB": 25 - }, - "resume_data": "Harvey Dent\nMachine Learning Engineer321 Gotham St\nGotham, USA\n\u2642\u00b6obile-alt(123) 456-7891\n/envel\u2322pehdent@email.com\n/linkedin-inhdent\n/githubhdent\nProfessional Summary\nMachine Learning Engineer with 5 years of experience in designing, building, and deploying predictive\nmodels and deep learning systems. Proficient in Python, TensorFlow, PyTorch, and Scikit-learn.\nCurrently leading a team of AI engineers at OpenAI.\nSkills\nPython, R, TensorFlow, PyTorch, Scikit-learn, Keras, SQL, NoSQL, Git, Docker,\nKubernetes, Agile and Scrum, Statistics, Data visualization, Deep Learning, Natural\nLanguage Processing\nExperience\n2021 \u2013\nPresentMachine Learning Engineer ,OpenAI , San Francisco, USA\nLeading a team of AI engineers. Designed and implemented deep learning models for natural\nlanguage processing tasks. Improved the efficiency of model training and data processing\npipelines. Published several research papers in top-tier AI conferences.\n2018 \u2013 2021 Data Scientist ,Uber , San Francisco, USA\nDeveloped and deployed machine learning models to improve the efficiency of ride allocation\nalgorithms. Utilized TensorFlow and PyTorch for developing predictive models. Analyzed\nand visualized large data sets to drive business strategies.\n2016 \u2013 2018 Junior Data Analyst ,Facebook , Menlo Park, USA\nAnalyzed and visualized large datasets using Python and R. Assisted in the development of\nmachine learning models for user behavior prediction. Conducted A/B testing and provided\ninsights to the product team.\nEducation\n2014 \u2013 2016 Master of Science in Computer Science, Specialization in AI ,MIT, Cambridge,\nUSA\n2010 \u2013 2014 Bachelor of Science in Computer Science ,UC San Diego , San Diego, USA\nProjects\n2021 \u2013\nPresentPersonal Project: Predictive Stock Trading Model\nDeveloped a predictive stock trading model using deep learning and time series analysis.\nUsed PyTorch for model development and Docker for deployment. The model predicts stock\nprices with a high degree of accuracy and automates trading decisions.", - "tri_grams": "[Machine Learning Engineer321, Learning Engineer321 Gotham, Engineer321 Gotham St, Machine Learning Engineer, years of experience, experience in designing, building and deploying, models and deep, deep learning systems, learning systems Proficient, Proficient in Python, Python TensorFlow PyTorch, PyTorch and Scikitlearn, leading a team, team of AI, engineers at OpenAI, Python R TensorFlow, R TensorFlow PyTorch, TensorFlow PyTorch Scikitlearn, PyTorch Scikitlearn Keras, Scikitlearn Keras SQL, Keras SQL NoSQL, SQL NoSQL Git, NoSQL Git Docker, Agile and Scrum, Scrum Statistics Data, Statistics Data visualization, Data visualization Deep, visualization Deep Learning, Deep Learning Natural, PresentMachine Learning Engineer, Learning Engineer OpenAI, San Francisco USA, Leading a team, team of AI, AI engineers Designed, Designed and implemented, implemented deep learning, deep learning models, models for natural, language processing tasks, processing tasks Improved, Improved the efficiency, efficiency of model, training and data, Published several research, papers in toptier, toptier AI conferences, Data Scientist Uber, San Francisco USA, Developed and deployed, deployed machine learning, machine learning models, models to improve, improve the efficiency, efficiency of ride, algorithms Utilized TensorFlow, TensorFlow and PyTorch, PyTorch for developing, developing predictive models, predictive models Analyzed, visualized large data, large data sets, sets to drive, drive business strategies, Junior Data Analyst, Data Analyst Facebook, Menlo Park USA, Analyzed and visualized, visualized large datasets, datasets using Python, Python and R, machine learning models, models for user, user behavior prediction, behavior prediction Conducted, testing and provided, Master of Science, Science in Computer, Computer Science Specialization, Specialization in AI, AI MIT Cambridge, Bachelor of Science, Science in Computer, Computer Science UC, Science UC San, UC San Diego, San Diego USA, PresentPersonal Project Predictive, Project Predictive Stock, Predictive Stock Trading, Stock Trading Model, Developed a predictive, predictive stock trading, stock trading model, model using deep, learning and time, time series analysis, PyTorch for model, development and Docker, Docker for deployment, deployment The model, model predicts stock, degree of accuracy, accuracy and automates, automates trading decisions]", - "unique_id": "5dde9183-5af4-4107-91e1-7b2d66fec490", - "years": [] -} \ No newline at end of file diff --git a/Data/Processed/Resumes/Resume-john_doe.pdf817fb796-14b9-4819-a224-7eed66b3ce04.json b/Data/Processed/Resumes/Resume-john_doe.pdf817fb796-14b9-4819-a224-7eed66b3ce04.json deleted file mode 100644 index 2ffdfdfb..00000000 --- a/Data/Processed/Resumes/Resume-john_doe.pdf817fb796-14b9-4819-a224-7eed66b3ce04.json +++ /dev/null @@ -1,289 +0,0 @@ -{ - "bi_grams": "[JOHN DOE, Main St, LinkedIn linkedincom, GitHub githubcom, PROFESSIONAL SUMMARY, Highly skilled, Stack Developer, Angular development, designing building, building testing, maintaining web, web applications, applications Proficient, technologies including, including Java, Java Spring, Spring Boot, Boot Angular, Angular HTML5, HTML5 CSS3, SQL Exceptional, Exceptional ability, selfdirect Committed, providing highquality, highquality results, little supervision, \u2022Spring Boot, Boot Spring, Spring MVC, \u2022Angular versions, \u2022JavaScript TypeScript, TypeScript HTML5, HTML5 CSS3, \u2022RESTful APIs, NoSQL databases, databases MySQL, MySQL MongoDB, WORK EXPERIENCE, Stack Java, Java Developer, ABC Company, Company Inc, USA June, \u2022Developed scalable, scalable robust, maintainable enterpriselevel, enterpriselevel applications, \u2022Used Angular, developing dynamic, responsive web, web frontends, frontends improving, improving user, user experience, \u2022Integrated applications, MongoDB databases, retrieve data, data efficiently, Agile development, development team, deliver highquality, highquality software, \u2022Created RESTful, RESTful services, thirdparty applications, \u2022Wrote unit, unit tests, robust testing, application components, Software Developer, XYZ Solutions, USA July, complete software, software development, development life, life cycle, requirement analysis, \u2022Implemented business, business logic, enhanced user, user interface, maintained SQL, NoSQL databases, databases implementing, implementing complex, complex queries, \u2022Utilized Git, version control, team members, troubleshooting software, software debugging, system enhancements, Computer Science, State University, \u2022Oracle Certified, Certified Professional, Professional Java, Java SE, \u2022Certified Angular, Angular Developer]", - "clean_data": "JOHN DOE\n123 Main St Anywhere USA \nLinkedIn linkedincom/in/johndoe GitHub githubcom/johndoe\nPROFESSIONAL SUMMARY\nHighly skilled Full Stack Developer with over 5 years of experience in Java and Angular development\nspecializing in designing building testing and maintaining web applications Proficient in an assortment\nof technologies including Java Spring Boot Angular HTML5 CSS3 and SQL Exceptional ability to\nwork in a team and selfdirect Committed to providing highquality results with little supervision\nSKILLS\n\u2022Java and J2EE\n\u2022Spring Boot Spring MVC and Hibernate\n\u2022Angular versions 2+\n\u2022JavaScript TypeScript HTML5 CSS3 and Bootstrap\n\u2022RESTful APIs\n\u2022SQL and NoSQL databases MySQL MongoDB\n\u2022Agile and Scrum\n\u2022Git and GitHub\n\u2022Junit and Mockito\n\u2022Docker\nWORK EXPERIENCE\nFull Stack Java Developer ABC Company Inc Anywhere USA June 2018 Present\n\u2022Developed scalable robust and maintainable enterpriselevel applications using Java and Spring\nBoot\n\u2022Used Angular for developing dynamic and responsive web frontends improving user experience\nby 30\n\u2022Integrated applications with MySQL and MongoDB databases to store and retrieve data efficiently\n\u2022Collaborated in an Agile development team to deliver highquality software every sprint\n\u2022Created RESTful services and APIs for frontend and thirdparty applications\n\u2022Wrote unit tests using Junit and Mockito for robust testing of application components\nSoftware Developer XYZ Solutions Anywhere USA July 2016 June 2018\n\u2022Participated in the complete software development life cycle from requirement analysis to deploy\nment\n\u2022Implemented business logic using Java and enhanced user interface using Angular\n\u2022Developed and maintained SQL and NoSQL databases implementing complex queries for business\nneeds\n\u2022Utilized Git for version control and collaborated with team members via GitHub\n\u2022Assisted in troubleshooting software debugging and system enhancements\n1 EDUCATION\nBachelor of Science in Computer Science State University Anywhere USA May 2016\nCERTIFICATIONS\n\u2022Oracle Certified Professional Java SE 8 Programmer\n\u2022Certified Angular Developer\n2", - "emails": [ - "john.doe@email.com" - ], - "entities": [ - "TypeScript", - "SQL Exceptional", - "Bachelor of Science in Computer Science State University", - "Committed", - "Software Developer XYZ Solutions", - "SQL", - "CERTIFICATIONS", - "LinkedIn", - "Java Spring Boot Angular HTML5 CSS3", - "Mockito", - "GitHub", - "Boot Spring", - "\u2022SQL", - "NoSQL" - ], - "experience": "", - "extracted_keywords": [ - "JOHN", - "DOE", - "Main", - "St", - "Anywhere", - "USA", - "LinkedIn", - "linkedincom", - "GitHub", - "githubcom", - "johndoe", - "PROFESSIONAL", - "SUMMARY", - "Stack", - "Developer", - "years", - "experience", - "Java", - "Angular", - "development", - "building", - "testing", - "web", - "applications", - "assortment", - "technologies", - "Java", - "Spring", - "Boot", - "Angular", - "HTML5", - "CSS3", - "SQL", - "ability", - "team", - "highquality", - "results", - "supervision", - "SKILLS", - "\u2022Java", - "J2EE", - "Boot", - "Spring", - "MVC", - "Hibernate", - "\u2022Angular", - "versions", - "TypeScript", - "HTML5", - "CSS3", - "Bootstrap", - "APIs", - "\u2022SQL", - "NoSQL", - "MySQL", - "MongoDB", - "\u2022Agile", - "Scrum", - "\u2022Git", - "GitHub", - "\u2022Junit", - "Mockito", - "\u2022Docker", - "EXPERIENCE", - "Stack", - "Java", - "Developer", - "ABC", - "Company", - "Inc", - "Anywhere", - "USA", - "June", - "Present", - "enterpriselevel", - "applications", - "Java", - "Spring", - "Boot", - "Angular", - "web", - "frontends", - "user", - "experience", - "applications", - "MySQL", - "MongoDB", - "databases", - "data", - "Agile", - "development", - "team", - "highquality", - "software", - "sprint", - "services", - "APIs", - "frontend", - "thirdparty", - "applications", - "\u2022Wrote", - "unit", - "tests", - "Junit", - "Mockito", - "testing", - "application", - "components", - "Software", - "Developer", - "XYZ", - "Solutions", - "Anywhere", - "USA", - "July", - "June", - "software", - "development", - "life", - "cycle", - "analysis", - "ment", - "business", - "logic", - "Java", - "user", - "interface", - "Angular", - "SQL", - "NoSQL", - "queries", - "business", - "needs", - "Git", - "version", - "control", - "team", - "members", - "GitHub", - "software", - "debugging", - "system", - "enhancements", - "EDUCATION", - "Bachelor", - "Science", - "Computer", - "Science", - "State", - "University", - "Anywhere", - "USA", - "May", - "CERTIFICATIONS", - "\u2022Oracle", - "Professional", - "Java", - "SE", - "Programmer", - "Angular", - "Developer" - ], - "keyterms": [ - [ - "HTML5 CSS3", - 0.15659926916055633 - ], - [ - "USA", - 0.05051062063174647 - ], - [ - "Java", - 0.0408365220696437 - ], - [ - "Angular", - 0.03796650660471361 - ], - [ - "application", - 0.03774315977767331 - ], - [ - "GitHub", - 0.03518440373889579 - ], - [ - "experience", - 0.03217193993366211 - ], - [ - "Spring", - 0.027840505763088406 - ], - [ - "development", - 0.027041932986194655 - ], - [ - "web", - 0.02066359533859646 - ], - [ - "testing", - 0.019103827923984563 - ], - [ - "Boot", - 0.018846596952617416 - ], - [ - "team", - 0.018417794514478142 - ], - [ - "developer", - 0.017014280950798117 - ], - [ - "highquality", - 0.014520497315477006 - ], - [ - "api", - 0.011352669917397118 - ], - [ - "version", - 0.011098703141591258 - ], - [ - "mysql", - 0.01090181270714155 - ], - [ - "MongoDB", - 0.010759973199125123 - ], - [ - "NoSQL", - 0.010192839972429344 - ] - ], - "name": [], - "phones": [], - "pos_frequencies": { - "ADJ": 20, - "ADP": 27, - "ADV": 2, - "CCONJ": 24, - "DET": 5, - "NOUN": 72, - "NUM": 11, - "PART": 4, - "PROPN": 89, - "SPACE": 48, - "SYM": 5, - "VERB": 37 - }, - "resume_data": "JOHN DOE\n123 Main St, Anywhere, USA \u2014 (123) 456-7890 \u2014 john.doe@email.com\nLinkedIn: linkedin.com/in/johndoe \u2014 GitHub: github.com/johndoe\nPROFESSIONAL SUMMARY\nHighly skilled Full Stack Developer with over 5 years of experience in Java and Angular development,\nspecializing in designing, building, testing, and maintaining web applications. Proficient in an assortment\nof technologies, including Java, Spring Boot, Angular, HTML5, CSS3, and SQL. Exceptional ability to\nwork in a team and self-direct. Committed to providing high-quality results with little supervision.\nSKILLS\n\u2022Java and J2EE\n\u2022Spring Boot, Spring MVC, and Hibernate\n\u2022Angular (versions 2+)\n\u2022JavaScript, TypeScript, HTML5, CSS3, and Bootstrap\n\u2022RESTful APIs\n\u2022SQL and NoSQL databases (MySQL, MongoDB)\n\u2022Agile and Scrum\n\u2022Git and GitHub\n\u2022Junit and Mockito\n\u2022Docker\nWORK EXPERIENCE\nFull Stack Java Developer , ABC Company, Inc., Anywhere, USA, June 2018 - Present\n\u2022Developed scalable, robust, and maintainable enterprise-level applications using Java and Spring\nBoot\n\u2022Used Angular for developing dynamic and responsive web front-ends, improving user experience\nby 30\n\u2022Integrated applications with MySQL and MongoDB databases to store and retrieve data efficiently\n\u2022Collaborated in an Agile development team to deliver high-quality software every sprint\n\u2022Created RESTful services and APIs for frontend and third-party applications\n\u2022Wrote unit tests using Junit and Mockito for robust testing of application components\nSoftware Developer , XYZ Solutions, Anywhere, USA, July 2016 - June 2018\n\u2022Participated in the complete software development life cycle from requirement analysis to deploy-\nment\n\u2022Implemented business logic using Java and enhanced user interface using Angular\n\u2022Developed and maintained SQL and NoSQL databases, implementing complex queries for business\nneeds\n\u2022Utilized Git for version control and collaborated with team members via GitHub\n\u2022Assisted in troubleshooting, software debugging, and system enhancements\n1 EDUCATION\nBachelor of Science in Computer Science , State University, Anywhere, USA, May 2016\nCERTIFICATIONS\n\u2022Oracle Certified Professional, Java SE 8 Programmer\n\u2022Certified Angular Developer\n2", - "tri_grams": "[St Anywhere USA, skilled Full Stack, years of experience, experience in Java, Java and Angular, specializing in designing, designing building testing, testing and maintaining, maintaining web applications, web applications Proficient, technologies including Java, including Java Spring, Java Spring Boot, Spring Boot Angular, Boot Angular HTML5, Angular HTML5 CSS3, CSS3 and SQL, SQL Exceptional ability, team and selfdirect, Committed to providing, providing highquality results, results with little, \u2022Java and J2EE, \u2022Spring Boot Spring, Boot Spring MVC, MVC and Hibernate, \u2022JavaScript TypeScript HTML5, TypeScript HTML5 CSS3, CSS3 and Bootstrap, \u2022SQL and NoSQL, NoSQL databases MySQL, databases MySQL MongoDB, \u2022Agile and Scrum, \u2022Git and GitHub, \u2022Junit and Mockito, Stack Java Developer, ABC Company Inc, Inc Anywhere USA, \u2022Developed scalable robust, robust and maintainable, maintainable enterpriselevel applications, applications using Java, Java and Spring, Angular for developing, dynamic and responsive, responsive web frontends, web frontends improving, frontends improving user, improving user experience, applications with MySQL, MySQL and MongoDB, databases to store, store and retrieve, retrieve data efficiently, Agile development team, team to deliver, deliver highquality software, software every sprint, \u2022Created RESTful services, services and APIs, APIs for frontend, frontend and thirdparty, \u2022Wrote unit tests, tests using Junit, Junit and Mockito, Mockito for robust, testing of application, Solutions Anywhere USA, complete software development, software development life, development life cycle, cycle from requirement, analysis to deploy, \u2022Implemented business logic, logic using Java, Java and enhanced, enhanced user interface, interface using Angular, \u2022Developed and maintained, SQL and NoSQL, NoSQL databases implementing, databases implementing complex, implementing complex queries, queries for business, Git for version, control and collaborated, collaborated with team, members via GitHub, \u2022Assisted in troubleshooting, troubleshooting software debugging, debugging and system, Bachelor of Science, Science in Computer, University Anywhere USA, \u2022Oracle Certified Professional, Certified Professional Java, Professional Java SE, \u2022Certified Angular Developer]", - "unique_id": "817fb796-14b9-4819-a224-7eed66b3ce04", - "years": [] -} \ No newline at end of file diff --git a/resume_matcher/__init__.py b/resume_matcher/__init__.py new file mode 100644 index 00000000..e69de29b diff --git a/resume_matcher/dataextractor/DataExtractor.py b/resume_matcher/dataextractor/DataExtractor.py new file mode 100644 index 00000000..c4c7cc23 --- /dev/null +++ b/resume_matcher/dataextractor/DataExtractor.py @@ -0,0 +1,221 @@ +import re +import urllib + +import spacy + +from resume_matcher.dataextractor.TextCleaner import TextCleaner + +# Load the English model +nlp = spacy.load("en_core_web_md") + +RESUME_SECTIONS = [ + "Contact Information", + "Objective", + "Summary", + "Education", + "Experience", + "Skills", + "Projects", + "Certifications", + "Licenses", + "Awards", + "Honors", + "Publications", + "References", + "Technical Skills", + "Computer Skills", + "Programming Languages", + "Software Skills", + "Soft Skills", + "Language Skills", + "Professional Skills", + "Transferable Skills", + "Work Experience", + "Professional Experience", + "Employment History", + "Internship Experience", + "Volunteer Experience", + "Leadership Experience", + "Research Experience", + "Teaching Experience", +] + + +class DataExtractor: + """ + A class for extracting various types of data from text. + """ + + def __init__(self, raw_text: str): + """ + Initialize the DataExtractor object. + + Args: + raw_text (str): The raw input text. + """ + + self.text = raw_text + self.clean_text = TextCleaner.clean_text(self.text) + self.doc = nlp(self.clean_text) + + def extract_links(self): + """ + Find links of any type in a given string. + + Args: + text (str): The string to search for links. + + Returns: + list: A list containing all the found links. + """ + link_pattern = r"\b(?:https?://|www\.)\S+\b" + links = re.findall(link_pattern, self.text) + return links + + def extract_links_extended(self): + """ + Extract links of all kinds (HTTP, HTTPS, FTP, email, www.linkedin.com, + and github.com/user_name) from a webpage. + + Args: + url (str): The URL of the webpage. + + Returns: + list: A list containing all the extracted links. + """ + links = [] + try: + response = urllib.request.urlopen(self.text) + html_content = response.read().decode("utf-8") + pattern = r'href=[\'"]?([^\'" >]+)' + raw_links = re.findall(pattern, html_content) + for link in raw_links: + if link.startswith( + ( + "http://", + "https://", + "ftp://", + "mailto:", + "www.linkedin.com", + "github.com/", + "twitter.com", + ) + ): + links.append(link) + except Exception as e: + print(f"Error extracting links: {str(e)}") + return links + + def extract_names(self): + """Extracts and returns a list of names from the given + text using spaCy's named entity recognition. + + Args: + text (str): The text to extract names from. + + Returns: + list: A list of strings representing the names extracted from the text. + """ + names = [ent.text for ent in self.doc.ents if ent.label_ == "PERSON"] + return names + + def extract_emails(self): + """ + Extract email addresses from a given string. + + Args: + text (str): The string from which to extract email addresses. + + Returns: + list: A list containing all the extracted email addresses. + """ + email_pattern = r"\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}\b" + emails = re.findall(email_pattern, self.text) + return emails + + def extract_phone_numbers(self): + """ + Extract phone numbers from a given string. + + Args: + text (str): The string from which to extract phone numbers. + + Returns: + list: A list containing all the extracted phone numbers. + """ + phone_number_pattern = ( + r"^(\+\d{1,3})?[-.\s]?\(?\d{3}\)?[-.\s]?\d{3}[-.\s]?\d{4}$" + ) + phone_numbers = re.findall(phone_number_pattern, self.text) + return phone_numbers + + def extract_experience(self): + """ + Extract experience from a given string. It does so by using the Spacy module. + + Args: + text (str): The string from which to extract experience. + + Returns: + str: A string containing all the extracted experience. + """ + experience_section = [] + in_experience_section = False + + for token in self.doc: + if token.text in RESUME_SECTIONS: + if token.text == "Experience" or "EXPERIENCE" or "experience": + in_experience_section = True + else: + in_experience_section = False + + if in_experience_section: + experience_section.append(token.text) + + return " ".join(experience_section) + + def extract_position_year(self): + """ + Extract position and year from a given string. + + Args: + text (str): The string from which to extract position and year. + + Returns: + list: A list containing the extracted position and year. + """ + position_year_search_pattern = ( + r"(\b\w+\b\s+\b\w+\b),\s+(\d{4})\s*-\s*(\d{4}|\bpresent\b)" + ) + position_year = re.findall(position_year_search_pattern, self.text) + return position_year + + def extract_particular_words(self): + """ + Extract nouns and proper nouns from the given text. + + Args: + text (str): The input text to extract nouns from. + + Returns: + list: A list of extracted nouns. + """ + pos_tags = ["NOUN", "PROPN"] + nouns = [token.text for token in self.doc if token.pos_ in pos_tags] + return nouns + + def extract_entities(self): + """ + Extract named entities of types 'GPE' (geopolitical entity) and 'ORG' (organization) from the given text. + + Args: + text (str): The input text to extract entities from. + + Returns: + list: A list of extracted entities. + """ + entity_labels = ["GPE", "ORG"] + entities = [ + token.text for token in self.doc.ents if token.label_ in entity_labels + ] + return list(set(entities)) diff --git a/resume_matcher/dataextractor/KeyTermExtractor.py b/resume_matcher/dataextractor/KeyTermExtractor.py new file mode 100644 index 00000000..2b51c1d3 --- /dev/null +++ b/resume_matcher/dataextractor/KeyTermExtractor.py @@ -0,0 +1,154 @@ +import spacy +import textacy +from textacy import extract + +# Load the English model +nlp = spacy.load("en_core_web_md") + +RESUME_SECTIONS = [ + "Contact Information", + "Objective", + "Summary", + "Education", + "Experience", + "Skills", + "Projects", + "Certifications", + "Licenses", + "Awards", + "Honors", + "Publications", + "References", + "Technical Skills", + "Computer Skills", + "Programming Languages", + "Software Skills", + "Soft Skills", + "Language Skills", + "Professional Skills", + "Transferable Skills", + "Work Experience", + "Professional Experience", + "Employment History", + "Internship Experience", + "Volunteer Experience", + "Leadership Experience", + "Research Experience", + "Teaching Experience", +] + +REGEX_PATTERNS = { + "email_pattern": r"\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}\b", + "phone_pattern": r"\(?\d{3}\)?[-.\s]?\d{3}[-.\s]?\d{4}", + "link_pattern": r"\b(?:https?://|www\.)\S+\b", +} + +READ_RESUME_FROM = "Data/Resumes/" +SAVE_DIRECTORY_RESUME = "Data/Processed/Resumes" + +READ_JOB_DESCRIPTION_FROM = "Data/JobDescription/" +SAVE_DIRECTORY_JOB_DESCRIPTION = "Data/Processed/JobDescription" + + +class KeytermExtractor: + """ + A class for extracting keyterms from a given text using various algorithms. + """ + + def __init__(self, raw_text: str, top_n_values: int = 20): + """ + Initialize the KeytermExtractor object. + + Args: + raw_text (str): The raw input text. + top_n_values (int): The number of top keyterms to extract. + """ + self.raw_text = raw_text + self.text_doc = textacy.make_spacy_doc(self.raw_text, lang="en_core_web_md") + self.top_n_values = top_n_values + + def get_keyterms_based_on_textrank(self): + """ + Extract keyterms using the TextRank algorithm. + + Returns: + List[str]: A list of top keyterms based on TextRank. + """ + return list( + extract.keyterms.textrank( + self.text_doc, normalize="lemma", topn=self.top_n_values + ) + ) + + def get_keyterms_based_on_sgrank(self): + """ + Extract keyterms using the SGRank algorithm. + + Returns: + List[str]: A list of top keyterms based on SGRank. + """ + return list( + extract.keyterms.sgrank( + self.text_doc, normalize="lemma", topn=self.top_n_values + ) + ) + + def get_keyterms_based_on_scake(self): + """ + Extract keyterms using the sCAKE algorithm. + + Returns: + List[str]: A list of top keyterms based on sCAKE. + """ + return list( + extract.keyterms.scake( + self.text_doc, normalize="lemma", topn=self.top_n_values + ) + ) + + def get_keyterms_based_on_yake(self): + """ + Extract keyterms using the YAKE algorithm. + + Returns: + List[str]: A list of top keyterms based on YAKE. + """ + return list( + extract.keyterms.yake( + self.text_doc, normalize="lemma", topn=self.top_n_values + ) + ) + + def bi_gramchunker(self): + """ + Chunk the text into bigrams. + + Returns: + List[str]: A list of bigrams. + """ + return list( + textacy.extract.basics.ngrams( + self.text_doc, + n=2, + filter_stops=True, + filter_nums=True, + filter_punct=True, + ) + ) + + def tri_gramchunker(self): + """ + Chunk the text into trigrams. + + Returns: + List[str]: A list of trigrams. + """ + return list( + textacy.extract.basics.ngrams( + self.text_doc, + n=3, + filter_stops=True, + filter_nums=True, + filter_punct=True, + ) + ) diff --git a/resume_matcher/dataextractor/TextCleaner.py b/resume_matcher/dataextractor/TextCleaner.py new file mode 100644 index 00000000..fef5d701 --- /dev/null +++ b/resume_matcher/dataextractor/TextCleaner.py @@ -0,0 +1,124 @@ +import re + +import spacy + +# Load the English model +nlp = spacy.load("en_core_web_md") + +RESUME_SECTIONS = [ + "Contact Information", + "Objective", + "Summary", + "Education", + "Experience", + "Skills", + "Projects", + "Certifications", + "Licenses", + "Awards", + "Honors", + "Publications", + "References", + "Technical Skills", + "Computer Skills", + "Programming Languages", + "Software Skills", + "Soft Skills", + "Language Skills", + "Professional Skills", + "Transferable Skills", + "Work Experience", + "Professional Experience", + "Employment History", + "Internship Experience", + "Volunteer Experience", + "Leadership Experience", + "Research Experience", + "Teaching Experience", +] + +REGEX_PATTERNS = { + "email_pattern": r"\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}\b", + "phone_pattern": r"\(?\d{3}\)?[-.\s]?\d{3}[-.\s]?\d{4}", + "link_pattern": r"\b(?:https?://|www\.)\S+\b", +} + +READ_RESUME_FROM = "Data/Resumes/" +SAVE_DIRECTORY_RESUME = "Data/Processed/Resumes" + +READ_JOB_DESCRIPTION_FROM = "Data/JobDescription/" +SAVE_DIRECTORY_JOB_DESCRIPTION = "Data/Processed/JobDescription" + + +class TextCleaner: + """ + A class for cleaning a text by removing specific patterns. + """ + + def remove_emails_links(text): + """ + Clean the input text by removing specific patterns. + + Args: + text (str): The input text to clean. + + Returns: + str: The cleaned text. + """ + for pattern in REGEX_PATTERNS: + text = re.sub(REGEX_PATTERNS[pattern], "", text) + return text + + def clean_text(text): + """ + Clean the input text by removing specific patterns. + + Args: + text (str): The input text to clean. + + Returns: + str: The cleaned text. + """ + text = TextCleaner.remove_emails_links(text) + doc = nlp(text) + for token in doc: + if token.pos_ == "PUNCT": + text = text.replace(token.text, "") + return str(text) + + def remove_stopwords(text): + """ + Clean the input text by removing stopwords. + + Args: + text (str): The input text to clean. + + Returns: + str: The cleaned text. + """ + doc = nlp(text) + for token in doc: + if token.is_stop: + text = text.replace(token.text, "") + return text + + +class CountFrequency: + def __init__(self, text): + self.text = text + self.doc = nlp(text) + + def count_frequency(self): + """ + Count the frequency of words in the input text. + + Returns: + dict: A dictionary with the words as keys and the frequency as values. + """ + pos_freq = {} + for token in self.doc: + if token.pos_ in pos_freq: + pos_freq[token.pos_] += 1 + else: + pos_freq[token.pos_] = 1 + return pos_freq diff --git a/resume_matcher/dataextractor/__init__.py b/resume_matcher/dataextractor/__init__.py new file mode 100644 index 00000000..e69de29b diff --git a/resume_matcher/main.py b/resume_matcher/main.py new file mode 100644 index 00000000..7f19e4b4 --- /dev/null +++ b/resume_matcher/main.py @@ -0,0 +1,39 @@ +import os + +from resume_matcher.run_first import run_first +from resume_matcher.scripts.get_score import get_score +from resume_matcher.scripts.logger import init_logging_config +from resume_matcher.scripts.utils import find_path, read_json + +init_logging_config() + +run_first() + +cwd = find_path("Resume-Matcher") + +PROCESSED_RESUMES_PATH = os.path.join(cwd, "Data", "Processed", "Resumes/") +PROCESSED_JOB_DESCRIPTIONS_PATH = os.path.join( + cwd, "Data", "Processed", "JobDescription/" +) + + +def get_filenames_from_dir(directory): + return [f for f in os.listdir(directory) if os.path.isfile(os.path.join(directory, f))] + + +def process_files(resume, job_description): + resume_dict = read_json(PROCESSED_RESUMES_PATH + resume) + job_dict = read_json(PROCESSED_JOB_DESCRIPTIONS_PATH + job_description) + resume_keywords = resume_dict["extracted_keywords"] + job_description_keywords = job_dict["extracted_keywords"] + + resume_string = " ".join(resume_keywords) + jd_string = " ".join(job_description_keywords) + final_result = get_score(resume_string, jd_string) + for r in final_result: + print(r.score) + print(f"Processing resume: {resume}") + print(f"Processing job description: {job_description}") + + + diff --git a/resume_matcher/run_first.py b/resume_matcher/run_first.py new file mode 100644 index 00000000..1cd88a12 --- /dev/null +++ b/resume_matcher/run_first.py @@ -0,0 +1,59 @@ +import logging +import os + +from tqdm import tqdm +from resume_matcher.scripts.processor import Processor +from resume_matcher.scripts.utils import get_filenames_from_dir, find_path + +cwd = find_path("Resume-Matcher") +RESUMES_PATH = os.path.join(cwd, "Data", "Resumes/") +JOB_DESCRIPTIONS_PATH = os.path.join(cwd, "Data", "JobDescription/") +PROCESSED_RESUMES_PATH = os.path.join(cwd, "Data", "Processed", "Resumes/") +PROCESSED_JOB_DESCRIPTIONS_PATH = os.path.join( + cwd, "Data", "Processed", "JobDescription/" +) + +logger = logging.getLogger(__name__) + + +def remove_old_files(files_path): + for filename in os.listdir(files_path): + try: + file_path = os.path.join(files_path, filename) + + if os.path.isfile(file_path): + os.remove(file_path) + except Exception as e: + logging.error(f"Error deleting {file_path}:\n{e}") + + logging.info("Deleted old files from " + files_path) + + +def process_files(data_path, processed_path, file_type): + print(f"Processing {file_type}s from {data_path}") + logging.info(f"Started to read from {data_path}") + try: + remove_old_files(processed_path) + file_names = get_filenames_from_dir(data_path) + logging.info(f"Reading from {data_path} is now complete.") + except: + logging.error(f"There are no {file_type}s present in the specified folder.") + logging.error("Exiting from the program.") + logging.error( + f"Please add {file_type}s in the {data_path} folder and try again." + ) + exit(1) + + logging.info(f"Started parsing the {file_type}s.") + for file in tqdm(file_names): + processor_object = Processor(file, file_type) + success = processor_object.process() + print(f"Processing of {file_type}s is now complete.") + logging.info(f"Parsing of the {file_type}s is now complete.") + + +def run_first(): + process_files(RESUMES_PATH, PROCESSED_RESUMES_PATH, "resume") + process_files( + JOB_DESCRIPTIONS_PATH, PROCESSED_JOB_DESCRIPTIONS_PATH, "job_description" + ) diff --git a/resume_matcher/scripts/__init__.py b/resume_matcher/scripts/__init__.py new file mode 100644 index 00000000..e69de29b diff --git a/resume_matcher/scripts/get_score.py b/resume_matcher/scripts/get_score.py new file mode 100644 index 00000000..3466b6ff --- /dev/null +++ b/resume_matcher/scripts/get_score.py @@ -0,0 +1,77 @@ +import logging +import os +from typing import List + +from qdrant_client import QdrantClient + +from resume_matcher.scripts.utils import find_path, read_json + +# Get the logger +logger = logging.getLogger(__name__) + +# Set the logging level +logger.setLevel(logging.INFO) + + +cwd = find_path("Resume-Matcher") +READ_RESUME_FROM = os.path.join(cwd, "Data", "Processed", "Resumes/") +READ_JOB_DESCRIPTION_FROM = os.path.join(cwd, "Data", "Processed", "JobDescription/") + + +def get_score(resume_string, job_description_string): + """ + The function `get_score` uses QdrantClient to calculate the similarity score between a resume and a + job description. + + Args: + resume_string: The `resume_string` parameter is a string containing the text of a resume. It + represents the content of a resume that you want to compare with a job description. + job_description_string: The `get_score` function you provided seems to be using a QdrantClient to + calculate the similarity score between a resume and a job description. The function takes in two + parameters: `resume_string` and `job_description_string`, where `resume_string` is the text content + of the resume and + + Returns: + The function `get_score` returns the search result obtained by querying a QdrantClient with the + job description string against the resume string provided. + """ + logger.info("Started getting similarity score") + + documents: List[str] = [resume_string] + client = QdrantClient(":memory:") + client.set_model("BAAI/bge-base-en") + + client.add( + collection_name="demo_collection", + documents=documents, + ) + + search_result = client.query( + collection_name="demo_collection", query_text=job_description_string + ) + logger.info("Finished getting similarity score") + return search_result + + +def custom_test(): + # To give your custom resume use this code + resume_dict = read_json( + READ_RESUME_FROM + + "resume_barry_allen_fe.pdf44a91b3b-b553-4765-b6b8-bfe26135f87b.json" + ) + job_dict = read_json( + READ_JOB_DESCRIPTION_FROM + + "job_description_job_desc_front_end_engineer.pdf947c72ae-7faf-45fa-86a4-92db51c07b45.json" + ) + resume_keywords = resume_dict["extracted_keywords"] + job_description_keywords = job_dict["extracted_keywords"] + + resume_string = " ".join(resume_keywords) + jd_string = " ".join(job_description_keywords) + final_result = get_score(resume_string, jd_string) + for r in final_result: + print(r.score) + + +if __name__ == "__main__": + custom_test() diff --git a/resume_matcher/scripts/logger.py b/resume_matcher/scripts/logger.py new file mode 100644 index 00000000..25d01d79 --- /dev/null +++ b/resume_matcher/scripts/logger.py @@ -0,0 +1,148 @@ +import logging + + +def get_handlers( + filename="app.log", mode="w", file_level=logging.DEBUG, stderr_level=logging.DEBUG +): + """ + The function `get_handlers` returns a stream handler and a file handler with specified logging + levels and formatters. + + Args: + stderr_level: + file_level: + mode: + filename: The `filename` parameter is the name of the log file where the log messages will be + written. In this case, the default filename is "app.log". Defaults to app.log + mode: The `mode` parameter in the `get_handlers` function specifies the mode in which the file + should be opened. In this case, the default mode is set to "w", which stands for write mode. This + means that if the file already exists, it will be truncated (i.e., its. Defaults to w + file_level: The `file_level` parameter in the `get_handlers` function is used to specify the + logging level for the file handler. In this case, it is set to `logging.DEBUG`, which means that the + file handler will log all messages at the DEBUG level and above. + stderr_level: The `stderr_level` parameter in the `get_handlers` function is used to specify the + logging level for the StreamHandler that outputs log messages to the standard error stream (stderr). + This level determines which log messages will be processed and output by the StreamHandler. + + Returns: + The `get_handlers` function returns two logging handlers: `stderr_handler` which is a + StreamHandler for logging to stderr, and `file_handler` which is a FileHandler for logging to a file + specified by the `filename` parameter. + """ + # Stream handler + stderr_handler = logging.StreamHandler() + stderr_handler.setLevel(stderr_level) + stderr_handler.setFormatter(CustomFormatter()) + + # File handler + file_handler = logging.FileHandler(filename, mode=mode) + file_handler.setLevel(file_level) + file_handler.setFormatter(CustomFormatter(True)) + + # TODO: Add RotatingFileHandler + + return stderr_handler, file_handler + + +class CustomFormatter(logging.Formatter): + """ + A custom log formatter that adds color to log messages based on the log level. + + Args: + file (bool): Indicates whether the log is being written to a file. Default is False. + + Attributes: + FORMATS (dict): A dictionary mapping log levels to colorized log message formats. + + Methods: + format(record): Formats the log record with the appropriate colorized log message format. + + """ + + def __init__(self, file=False): + """ + This function initializes logging formats with different colors and styles based on the log + level. + + Args: + file: The `file` parameter in the `__init__` method is a boolean flag that determines whether + the logging output should be colored or not. If `file` is `True`, the colors will not be applied + to the log messages. Defaults to False + """ + super().__init__() + yellow = "\x1b[36;10m" if not file else "" + blue = "\x1b[35;10m" if not file else "" + green = "\x1b[32;10m" if not file else "" + red = "\x1b[31;10m" if not file else "" + bold_red = "\x1b[31;1m" if not file else "" + reset = "\x1b[0m" if not file else "" + log = "%(asctime)s (%(filename)s:%(lineno)d) - %(levelname)s: " + msg = reset + "%(message)s" + + self.FORMATS = { + logging.DEBUG: blue + log + msg, + logging.INFO: green + log + msg, + logging.WARNING: yellow + log + msg, + logging.ERROR: red + log + msg, + logging.CRITICAL: bold_red + log + msg, + } + + def format(self, record): + """ + Formats the log record with the appropriate colorized log message format. + + Args: + record (LogRecord): The log record to be formatted. + + Returns: + str: The formatted log message. + + """ + log_fmt = self.FORMATS.get(record.levelno) + formatter = logging.Formatter(log_fmt) + return formatter.format(record) + + +def init_logging_config( + basic_log_level=logging.INFO, + filename="app.log", + mode="w", + file_level=logging.DEBUG, + stderr_level=logging.DEBUG, +): + """ + The function `init_logging_config` initializes logging configuration in Python by setting basic log + level, configuring handlers, and adding them to the logger. + + Args: + basic_log_level: The `basic_log_level` parameter is used to set the logging level for the root + logger. In this function, it is set to `logging.INFO` by default, which means that log messages with + severity level INFO or higher will be processed. + filename: The `filename` parameter is a string that specifies the name of the log file where the + logs will be written. In the `init_logging_config` function you provided, the default value for + `filename` is "app.log". This means that if no filename is provided when calling the function, logs. + Defaults to app.log + mode: The `mode` parameter in the `init_logging_config` function specifies the mode in which the + log file will be opened. In this case, the default value is "w" which stands for write mode. This + means that the log file will be opened for writing, and if the file already exists. Defaults to w + file_level: The `file_level` parameter in the `init_logging_config` function is used to specify + the logging level for the file handler. This determines the severity level of log messages that will + be written to the log file specified by the `filename` parameter. In this case, the default value + for `file + stderr_level: The `stderr_level` parameter in the `init_logging_config` function is used to + specify the logging level for the stderr (standard error) handler. This handler is responsible for + directing log messages to the standard error stream. The logging level determines which severity of + log messages will be output to the stderr. + """ + + logger = logging.getLogger() + logger.setLevel(basic_log_level) + + # Get the handlers + stderr_handler, file_handler = get_handlers( + file_level=file_level, stderr_level=stderr_level, filename=filename, mode=mode + ) + + # Add the handlers + logger.addHandler(stderr_handler) + logger.addHandler(file_handler) diff --git a/resume_matcher/scripts/parser.py b/resume_matcher/scripts/parser.py new file mode 100644 index 00000000..2b91d3c3 --- /dev/null +++ b/resume_matcher/scripts/parser.py @@ -0,0 +1,55 @@ +from resume_matcher.dataextractor.DataExtractor import DataExtractor +from resume_matcher.dataextractor.KeyTermExtractor import KeytermExtractor +from resume_matcher.dataextractor.TextCleaner import TextCleaner, CountFrequency +from resume_matcher.scripts.utils import generate_unique_id + + +class ParseDocumentToJson: + def __init__(self, doc: str, doc_type: str): + self.years = None + self.phones = None + self.emails = None + self.experience = None + self.name = None + self.doc_data = doc + self.doc_type = doc_type + self.clean_data = TextCleaner.clean_text(self.doc_data) + self.entities = DataExtractor(self.clean_data).extract_entities() + self.key_words = DataExtractor(self.clean_data).extract_particular_words() + self.pos_frequencies = CountFrequency(self.clean_data).count_frequency() + self.keyterms = KeytermExtractor(self.clean_data).get_keyterms_based_on_sgrank() + self.bi_grams = KeytermExtractor(self.clean_data).bi_gramchunker() + self.tri_grams = KeytermExtractor(self.clean_data).tri_gramchunker() + if self.doc_type == "resume": + self.get_additional_data() + + def get_additional_data(self): + self.name = DataExtractor(self.clean_data[:30]).extract_names() + self.experience = DataExtractor(self.clean_data).extract_experience() + self.emails = DataExtractor(self.doc_data).extract_emails() + self.phones = DataExtractor(self.doc_data).extract_phone_numbers() + self.years = DataExtractor(self.clean_data).extract_position_year() + + def get_JSON(self) -> dict: + doc_dictionary = { + "unique_id": generate_unique_id(), + "doc_data": self.doc_data, + "clean_data": self.clean_data, + "entities": self.entities, + "extracted_keywords": self.key_words, + "keyterms": self.keyterms, + "bi_grams": str(self.bi_grams), + "tri_grams": str(self.tri_grams), + "pos_frequencies": self.pos_frequencies, + } + if self.doc_type == "resume": + doc_dictionary.update( + { + "name": self.name, + "experience": self.experience, + "emails": self.emails, + "phones": self.phones, + "years": self.years, + } + ) + return doc_dictionary diff --git a/resume_matcher/scripts/processor.py b/resume_matcher/scripts/processor.py new file mode 100644 index 00000000..7730b5df --- /dev/null +++ b/resume_matcher/scripts/processor.py @@ -0,0 +1,54 @@ +import json +import os +import os.path +import pathlib + +from .parser import ParseDocumentToJson +from .utils import read_single_pdf, find_path + +cwd = find_path("Resume-Matcher") + +READ_RESUME_FROM = os.path.join(cwd, "Data", "Resumes/") +SAVE_RESUME_TO = os.path.join(cwd, "Data", "Processed", "Resumes/") + +READ_JOB_DESCRIPTION_FROM = os.path.join(cwd, "Data", "JobDescription/") +SAVE_JOB_DESCRIPTION_TO = os.path.join(cwd, "Data", "Processed", "JobDescription/") + + +class Processor: + def __init__(self, input_file, file_type): + self.input_file = input_file + self.file_type = file_type + if file_type == "resume": + self.input_file_name = os.path.join(READ_RESUME_FROM + self.input_file) + elif file_type == "job_description": + self.input_file_name = os.path.join( + READ_JOB_DESCRIPTION_FROM + self.input_file + ) + + def process(self) -> bool: + try: + data_dict = self._read_data() + self._write_json_file(data_dict) + return True + except Exception as e: + print(f"An error occurred: {str(e)}") + return False + + def _read_data(self) -> dict: + data = read_single_pdf(self.input_file_name) + output = ParseDocumentToJson(data, self.file_type).get_JSON() + return output + + def _write_json_file(self, data_dict: dict): + file_name = str( + f"{self.file_type}_" + self.input_file + data_dict["unique_id"] + ".json" + ) + save_directory_name = None + if self.file_type == "resume": + save_directory_name = pathlib.Path(SAVE_RESUME_TO) / file_name + elif self.file_type == "job_description": + save_directory_name = pathlib.Path(SAVE_JOB_DESCRIPTION_TO) / file_name + json_object = json.dumps(data_dict, sort_keys=True, indent=14) + with open(save_directory_name, "w+") as outfile: + outfile.write(json_object) diff --git a/resume_matcher/scripts/utils.py b/resume_matcher/scripts/utils.py new file mode 100644 index 00000000..3dfc8fa8 --- /dev/null +++ b/resume_matcher/scripts/utils.py @@ -0,0 +1,147 @@ +import glob +import json +import logging +import os +import os.path +from uuid import uuid4 + +from pypdf import PdfReader + +logger = logging.getLogger(__name__) + + +def find_path(folder_name): + """ + The function `find_path` searches for a folder by name starting from the current directory and + traversing up the directory tree until the folder is found or the root directory is reached. + + Args: + folder_name: The `find_path` function you provided is designed to search for a folder by name + starting from the current working directory and moving up the directory tree until it finds the + folder or reaches the root directory. + + Returns: + The `find_path` function is designed to search for a folder with the given `folder_name` starting + from the current working directory (`os.getcwd()`). It iterates through the directory structure, + checking if the folder exists in the current directory or any of its parent directories. If the + folder is found, it returns the full path to that folder using `os.path.join(curr_dir, folder_name)` + """ + curr_dir = os.getcwd() + while True: + if folder_name in os.listdir(curr_dir): + return os.path.join(curr_dir, folder_name) + else: + parent_dir = os.path.dirname(curr_dir) + if parent_dir == "/": + break + curr_dir = parent_dir + raise ValueError(f"Folder '{folder_name}' not found.") + + +def read_json(path): + """ + The `read_json` function reads a JSON file from the specified path and returns its contents, handling + any exceptions that may occur during the process. + + Args: + path: The `path` parameter in the `read_doc` function is a string that represents the file path to + the JSON document that you want to read and load. This function reads the JSON data from the file + located at the specified path. + + Returns: + The function `read_doc(path)` reads a JSON file located at the specified `path`, and returns the + data loaded from the file. If there is an error reading the JSON file, it logs the error message and + returns an empty dictionary `{}`. + """ + with open(path) as f: + try: + data = json.load(f) + except Exception as e: + logger.error(f"Error reading JSON file: {e}") + data = {} + return data + + +def read_multiple_pdf(file_path: str) -> list: + """ + Read multiple PDF files from the specified file path and extract the text from each page. + + Args: + file_path (str): The directory path containing the PDF files. + + Returns: + list: A list containing the extracted text from each page of the PDF files. + """ + pdf_files = get_pdf_files(file_path) + output = [] + for file in pdf_files: + try: + with open(file, "rb") as f: + pdf_reader = PdfReader(f) + count = pdf_reader.getNumPages() + for i in range(count): + page = pdf_reader.getPage(i) + output.append(page.extractText()) + except Exception as e: + print(f"Error reading file '{file}': {str(e)}") + return output + + +def read_single_pdf(file_path: str) -> str: + """ + Read a single PDF file and extract the text from each page. + + Args: + file_path (str): The path of the PDF file. + + Returns: + list: A list containing the extracted text from each page of the PDF file. + """ + output = [] + try: + with open(file_path, "rb") as f: + pdf_reader = PdfReader(f) + count = len(pdf_reader.pages) + for i in range(count): + page = pdf_reader.pages[i] + output.append(page.extract_text()) + except Exception as e: + print(f"Error reading file '{file_path}': {str(e)}") + return str(" ".join(output)) + + +def get_pdf_files(file_path: str) -> list: + """ + Get a list of PDF files from the specified directory path. + + Args: + file_path (str): The directory path containing the PDF files. + + Returns: + list: A list of PDF file paths. + """ + pdf_files = [] + try: + pdf_files = glob.glob(os.path.join(file_path, "*.pdf")) + except Exception as e: + print(f"Error getting PDF files from '{file_path}': {str(e)}") + return pdf_files + + +def generate_unique_id(): + """ + Generate a unique ID and return it as a string. + + Returns: + str: A string with a unique ID. + """ + return str(uuid4()) + + +def get_filenames_from_dir(directory_path: str) -> list: + filenames = [ + f + for f in os.listdir(directory_path) + if os.path.isfile(os.path.join(directory_path, f)) and f != ".DS_Store" + ] + return filenames diff --git a/scripts/similarity/get_similarity_score.py b/scripts/similarity/get_similarity_score.py index 86b19494..3d2b3eae 100644 --- a/scripts/similarity/get_similarity_score.py +++ b/scripts/similarity/get_similarity_score.py @@ -2,7 +2,6 @@ import logging import os -import cohere import yaml from qdrant_client import QdrantClient, models from qdrant_client.http.models import Batch