PDF Chatbot (RAG)
Chat with any PDF: smart chunking, sentence embeddings, ChromaDB vector search and conversational memory.
- My role
- Developer
The problem
Long documents are hard to search. Retrieval-augmented generation lets you ask them questions instead.
What I built
- PDF text extraction and smart chunking.
- Hugging Face sentence embeddings stored in ChromaDB for vector search.
- Conversational memory and a Streamlit UI, with automatic CUDA/CPU handling in PyTorch.