A production-ready Retrieval-Augmented Generation (RAG) system with security features and FastAPI web service.
- Hybrid Search: Vector similarity + BM25 keyword search
- Reranking: Cross-encoder for precision
- Security: Input validation, PII detection, rate limiting
- API: FastAPI REST service with authentication
- Documentation: Auto-generated Swagger UI
- Python 3.11
- Sentence Transformers (embeddings)
- ChromaDB (vector database)
- Google Gemini (LLM)
- FastAPI (web framework)