DataGuardian AI

Project Story

Inspiration

As data ecosystems within enterprises grow increasingly complex, managing metadata, understanding data lineage, and enforcing governance have become massive bottlenecks. While robust metadata platforms like DataHub exist, interacting with them often requires deep technical knowledge and manual effort. We were inspired to build DataGuardian AI to bridge this gap: an intelligent workspace that leverages AI agents to autonomously discover, document, and govern data, making enterprise data intelligence accessible to everyone.

What it does

DataGuardian AI is an agent-driven data intelligence and governance workspace. It provides a seamless interface where specialized LangGraph AI agents interact with DataHub's metadata. Users can leverage these agents for tasks like automatic SQL generation, tracing data lineage, performing quality checks, and auto-generating documentation, all within a unified, modern web application.

How we built it

We adopted a decoupled, modern architecture to ensure scalability and smooth user experience:

• Frontend: We used Next.js 15 with TypeScript and Tailwind CSS to build a highly responsive and aesthetically pleasing user interface. Authentication is handled securely via GitHub OAuth using Auth.js.
• Backend: We built a high-performance REST API using Python 3.12 and FastAPI, which handles orchestration, authorization, and provider adapters.
• Database: PostgreSQL is used to persist application state and user data.
• Infrastructure: The entire stack is containerized using Docker and Docker Compose for easy local development, with GitHub Actions set up for Continuous Integration.

Challenges we ran into

One of the primary challenges was establishing a clean architectural boundary between the frontend interaction layer and the complex AI orchestration layer. Designing the backend to seamlessly coordinate LangGraph agents while maintaining DataHub as the absolute system of record for enterprise metadata required careful API contract design and rigorous testing.

Accomplishments that we're proud of

We are incredibly proud to have completed Phase 1 of our roadmap, establishing a rock-solid foundation. We successfully implemented a full-stack, decoupled application with secure OAuth, a robust database, and a fully containerized environment that can be spun up in minutes.

What we learned

Building DataGuardian AI reinforced the importance of choosing the right tech stack for the right job. Using Next.js allowed us to iterate rapidly on the user interface, while Python's rich AI ecosystem made FastAPI the perfect choice for our backend API. We also learned how critical well-defined API contracts are when preparing for complex agentic workflows.

Built With

• auth.js
• datahub
• docker
• fastapi
• langgraph
• next.js
• postgresql
• python
• react
• tailwind-css
• typescript