Project Overview
A demo-friendly retrieval pipeline that turns a static PDF into an interactive question-answering experience with document-aware responses.
Document Q&A (RAG)
Upload a PDF, index its contents, and ask questions against the document. The app uses sentence-transformers for embeddings, FAISS for retrieval, and a Hugging Face-hosted LLM for grounded answers.
Tech Stack
- → sentence-transformers embeddings
- → FAISS vector retrieval
- → Hugging Face-hosted LLM
- → PDF upload and Q&A workflow
Key Features
- ✓ Upload a PDF and extract searchable chunks automatically
- ✓ Retrieve the most relevant passages with FAISS similarity search
- ✓ Generate grounded answers from the uploaded document context
- ✓ Simple interface that works well as a portfolio demo
Project Focus
This project is designed to showcase practical RAG fundamentals: document ingestion, chunking, embedding generation, vector search, and answer synthesis from retrieved context.