RAG Policy copilot

Developed an intelligent document assistant that allows users to upload PDF documents and ask questions about their content using Retrieval-Augmented Generation (RAG).

**Key Features:**

- PDF ingestion, text extraction, and document indexing.

- Semantic search using embeddings and vector retrieval.

- Retrieval optimization using Reciprocal Rank Fusion (RRF).

- Context-aware responses generated by Large Language Models.

- Source citations with document names and page references.

- Evidence selection and answer validation to improve reliability.

- Interactive React chat interface with streaming responses.

- FastAPI backend and Docker-based deployment.

**Technologies:** Python, FastAPI, React, RAG, LLMs, Embeddings, Vector Search, Docker, REST APIs.

The system is designed to help businesses quickly retrieve information from internal documents while providing verifiable, source-grounded answers.