End-to-end Retrieval-Augmented Generation platform for document summarization and Q&A. Built solo as a capstone, March to May 2025.
What I built
- Secure FastAPI backend with JWT authentication for user, document, and API-key management.
- RAG pipeline with transformer-based embeddings + prompt engineering for document Q&A.
- User + admin dashboards via Streamlit with real-time usage analytics.
- Containerized with Docker Compose for scalable deployment.
Stack
Python · FastAPI · JWT · LangChain-style RAG · transformer embeddings · Streamlit · Docker Compose.
Why this is on the portfolio
Demonstrates the full-stack AI-shipping capability before LeoHydra. That capability is not a one-time fluke. Two months solo, deployed and demoable.
Try it at smartdoc-ai-user.streamlit.app.