A cloud-hosted AI application that allows users to upload documents and ask natural-language questions over their content using Retrieval-Augmented Generation (RAG).
Organizations and individuals often store knowledge in PDFs and documents that are difficult to search. This project demonstrates how to build a secure, scalable AI system that provides grounded answers over private data.
- Secure user authentication
- Document upload and async ingestion
- Vector search using embeddings
- AI-powered Q&A with citations
- Conversation history
See docs/architecture.md for system design and data flow.
- Backend: FastAPI (Python)
- AI: LLM + embeddings (RAG)
- Database: PostgreSQL + pgvector
- Storage: S3
- Queue: SQS
- Frontend: Next.js
- Cloud: AWS (ECS, RDS)
- IaC: Terraform