AI · Retrieval-Augmented Generation
Business Docs RAG Chatbot
Upload a business document. Ask a question. Retrieve the context behind the answer.
The Problem
Finding a specific answer in business documents can require repeated manual searching. This project connects document ingestion with retrieval-assisted question answering.
What I Built
Flask upload and chat endpoints with session-specific document ingestion.
Chroma collections that retrieve relevant document chunks for each session.
A LangChain pipeline with Groq inference, prompted to answer from retrieved context and acknowledge missing information.
Outcome
A document question-answering API with session-scoped retrieval and cached chains for follow-up queries.
Tech Stack
PythonFlaskLangChainGroqChromaSentence Transformers