About
Serverless RAG is a scalable document intelligence platform that leverages Retrieval-Augmented Generation to provide semantic search capabilities across enterprise data. It transforms static PDFs into a queryable knowledge base, allowing users to interact with their documents using natural language.
The system features a dual-stage architecture: a high-performance ingestion pipeline that encodes document chunks into vectors using Amazon Titan, and a real-time question-answering interface that synthesizes answers using Claude or GPT. It handles the complexity of vector synchronization across instances via an S3-backed layer, ensuring consistency in distributed environments.
Tools & process
- AWS Ecosystem — Orchestrated via EC2, S3, and ALB within a secure VPC isolation layer to ensure enterprise-grade reliability and security.
- Vector Intelligence — Utilizes FAISS for high-performance similarity search and Amazon Titan for state-of-the-at semantic embeddings.
- Full-Stack Implementation — Developed with a Python 3.11 backend and a Streamlit-powered frontend for rapid deployment and intuitive admin/client interfaces.