← All projects

RAG Research Chatbot

A chatbot for asking questions about your own research papers, built with retrieval-augmented generation.

Problem
Finding answers across many research papers takes time.
Approach
Papers are split into chunks and stored in a vector database for semantic search. A question-answering chain handles single questions, and a conversation chain keeps memory for follow-up questions. A small evaluation script checks answers against expected ones.
Interfaces
Command line and a Streamlit web app.
Tools
Python, LangChain, OpenAI, Streamlit