AI Retrieval-Augmented Generation (RAG) System project (AI-powered document Q&A)
Built an AI Retrieval-Augmented Generation (RAG) system to answer questions over text documents using semantic embeddings and a vector database. Implemented multilingual query support and low-latency retrieval to improve accuracy and responsiveness for end-user interactions. Conducted testing to validate retrieval correctness and pipeline reliability for document-based question answering. • Used Google Embedding Model to generate semantic embeddings for documents. • Stored and queried vectors in Supabase for efficient similarity search. • Integrated a Node.js backend to orchestrate the RAG pipeline via LangChain. • Added multilingual processing and performance optimizations for user experience.