RAG Chatbot

A chatbot that answers questions about any PDF document using Retrieval Augmented Generation (RAG).

Built as Project 4 of my AI portfolio.

What it does

• Upload any PDF document
• Ask questions about it in plain English
• Get accurate answers based on the document content

How it works

1. PDF is loaded and split into 500 word chunks
2. Each chunk is converted to vectors using sentence transformers
3. Vectors are stored in FAISS vector database
4. Your question is also converted to a vector
5. FAISS finds the 3 most similar chunks
6. TinyLlama reads those chunks and generates answer

Example

Upload your CV and ask: "What job positions match this resume?" The chatbot reads your CV and answers intelligently.

Built with

• Python 3.14
• LangChain
• FAISS — vector similarity search
• HuggingFace Sentence Transformers
• TinyLlama via Ollama
• Streamlit — web interface
• PyPDF — PDF reading

How to run

pip install langchain langchain-community langchain-text-splitters faiss-cpu sentence-transformers pypdf streamlit langchain-core

ollama pull tinyllama

streamlit run app.py