Skip to content
View sa-paul's full-sized avatar

Highlights

  • Pro

Block or report sa-paul

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
sa-paul/README.md

Welcome👋, I am Sayan Paul

Typing SVGTyping SVG

SWE Intern'25 @ Samsung R&D Bangalore • Ex-Research Intern'24 @ ACM IKDD UpLink IIT Kgp • B.Tech CSE @ NIT Durgapur'25
• Full Stack Developer • Deep Learning Enthusiast


Connect with me:

sayanPaulCSE sayanPaulCSE sayanPaulCSE


  • Worked on LLM-based ASR systems, including Gemini Flash and Pro, Qwen Audio, and Whisper, to mitigate gender bias in speech-to-text translation. Achieved a 51.04% reduction in bias by leveraging prompt engineering and employing evaluation metrics such as WER, AWG, and the Kruskal-Wallis test. Additionally, audited bias related to native region, age, and speech rate of speakers.

  • 🚀 Designed, Developed and Deployed a open-source browser extension, CF FetchCodes, for CodeForces, a competitive programming platform, to access all the accepted programs, written by CodeForces-Friends of a user, on a single click, using HTML, CSS and JavaScript, available in Chrome Web Store and Firefox Ad-ons Store.

  • ⚙️ Engineered DNS Based Ad Blocker, which blocks Ads in any device connected to a specific router, using Python and Socket Programming incorporating UDP protocols. Achieved a 50% reduction in ad traffic, improving network performance for over 150 connected devices on a specific router.

  • Developed a system for urban sound classification optimized for offline use on low-end devices like Raspberry Pi and Arduino. Leveraged custom CNN architectures, with feature extraction techniques including MFCC and spectrogram analysis, to ensure efficient and accurate performance on resource-constrained hardware.

  • 💬 Ask me about Computer Science Fundamentals, Competitive Programming and Deep Learning.

  • 👨‍💻 All of my Projects are available at Github.

  • 📄 Know about my Experiences at LinkedIn.


Languages and Tools:

c cplusplus python javascript html5 css3 mysql linux git tensorflow pytorch numpy pandas matplotlib google-cloud


My Most Used Languages:

sa-paul

Pinned Loading

  1. BlockADs BlockADs Public

    ADs Blocking Solution which blocks ADs on Web to all the devices, of any kind, connected to a specific network through routing devices.

    Python

  2. CF_FetchCodes CF_FetchCodes Public

    This is an extension to Chrome Browser to fetch accepted programs of a particular problem solved by user's CodeForces's Friends.

    JavaScript 1

  3. Audio_Classification_React_Native_App Audio_Classification_React_Native_App Public

    Classifies Urban Sound 8K in Android Application developed using react-native

  4. Train_Reservation_System Train_Reservation_System Public

    C++ project to implement Railway Reservation System

    C++

  5. Academic_Question_Answering-AQA-_KDD-2024 Academic_Question_Answering-AQA-_KDD-2024 Public

    A model capable of retrieving the most relevant papers to answer questions from various domains. Given professional questions and a pool of candidate papers, the objective is to retrieve the most r…