Back to projects
RAG Document Assistant
A document question-answering assistant that indexes uploaded PDFs with FAISS and answers questions through a FastAPI service and React interface.
Overview
RAG Document Assistant is a focused retrieval-augmented generation demo. Users can upload documents, build an in-memory FAISS index, and ask questions about the indexed PDF content.
Engineering Highlights
- Built a FastAPI backend with health, document listing, upload, and chat endpoints.
- Implemented PDF ingestion and in-memory FAISS retrieval for document-grounded answers.
- Added a React/Vite frontend and a deployment path using Hugging Face Spaces for the backend and Vercel for the frontend.
Tech Stack
Python, FastAPI, React, Vite, FAISS, PDF processing, Docker, and Gemini.
