Early-stage multi-agent code review platform (FastAPI backend, Next.js frontend, LangGraph + Groq). 22 commits in 5 days, typed Python, documented README, but no tests, no CI, unmerged conflicts in main files, and zero external adoption.

Indian Sign Language recognition pipeline with three neural networks (Expression CNN, Head Movement TCN, ISL Sentence BiLSTM) for real-time inference. Jupyter-based project with training notebooks, ONNX/PyTorch models, and entry-point Python script. No external adoption signals.

Early-stage floor planning web app (TypeScript + React + Vite) with interactive canvas, furniture catalog, and AI chat. Typed, structured, and documented, but thin README (Lovable template), no CI/tests beyond placeholder, nascent adoption.

Personal multi-agent RAG system with LangGraph orchestration, vector retrieval, and LLM-based verification. Typed Python with structured agents and FastAPI backend, but lacks README, tests, CI, and is under 2 weeks old with minimal commits.

Streamlit-based RAG system for querying PDFs via Milvus + Groq LLM. Has README, hybrid search implementation, and working dual CLI/web UIs, but untyped Python, no tests/CI, single day of development.

Jupyter Notebook animal face detection project with PyTorch CNN achieving 97% validation accuracy. No tests, CI, or license; educational scope with thin structure and limited adoption signal.

Single-week burst of a Jupyter Notebook-based rice classification demo, achieving 98.53% test accuracy on a Kaggle dataset using PyTorch. No tests, CI, or type hints; minimal architectural scope.

Personal portfolio README with no code artifacts—a professional profile page rather than a functional project. Contains claims of expertise (AI/LLM, FastAPI, RAG) but zero shipped code to validate them. 14 commits in ~10 months suggest minimal ongoing work.

Minimal one-shot dump: ~4 KB C++ order book implementation with 1 commit in 42 seconds (2026-02-27 05:48:50 to 05:49:32). Proof-of-concept code lacks tests, CI, and production rigor despite financial domain claims.

Jupyter Notebook collection on LangChain and AI without README, tests, CI, or documentation. No stars, newly created (16 days old), minimal structure observable.

Empty scaffold with minimal content (2 KB), created and last pushed same day with only 2 commits. No README, tests, CI, or documentation. Appears to be a one-time commit dump or abandoned setup.