LLM Document Analysis Pipeline
Python/FastAPI backend that answers natural-language questions over uploaded PDFs via a RAG pipeline with LangChain, Claude, Voyage embeddings, and a FAISS index, containerized with Docker. Hardened for public deployment with domain-scoped CORS, per-IP rate limiting, upload-size limits, and auth fixes; validated with a pytest unit and integration suite. Self-hosted: a static GitHub Pages frontend served through a Cloudflare Tunnel to the Dockerized API.
Python · FastAPI · LangChain · Claude · Voyage · FAISS · Docker