try to go around the location problem

.github/workflows/ci-shark-ai.yml

@@ -58,7 +58,7 @@ jobs:
           # from non default locations first. Installing the PyTorch CPU
           # wheels saves multiple minutes and a lot of bandwidth on runner setup.
           pip install --no-compile -r pytorch-cpu-requirements.txt
-          pip install --no-compile -r requirements.txt -e sharktank/ shortfin/
+          pip install --no-compile -r requirements.txt -e sharktank/
 
           # Install latest iree-tubrine.
           pip install --no-compile -f https://iree.dev/pip-release-links.html --src deps \
@@ -71,5 +71,8 @@ jobs:
             iree-base-compiler \
             iree-base-runtime
 
+      - name: Install shortfin
+        run: pip install --no-compile -e shortfin/
+
       - name: Run LLM Integration Tests
         run: pytest -v app_tests/integration_tests/llm --log-cli-level=INFO