Back to portfolio
Project case study

RAG Q&A — Document Intelligence & Search System

Retrieval-Augmented Generation system for intelligent document Q&A — reduces information retrieval time from hours to seconds.

RAGBM25Cross-EncoderCRAGLLM
RQ—
Project
Video walkthrough

Project demo embedded directly from YouTube.

Watch
Project story

Engineered a Hybrid Search (BM25 + Semantic) pipeline with a Cross-Encoder Re-ranker and Corrective RAG (CRAG) logic to improve retrieval accuracy by 45%.