Synapse Data Intelligence
Link to open source: https://github.com/humamaziz/Synapse-Data-Agent
The Problem: Enterprise database documentation is frequently outdated, missing, or lacks business context. This creates a "data downtime" crisis where technical metadata means nothing to business users, blocking data-driven decision-making.
Our Solution (Synapse): Synapse Data Intelligence is an AI-powered Data Dictionary Agent designed to completely automate enterprise data governance. Built for HackFest 2.0, our zero-friction platform seamlessly connects to relational databases to instantly extract complex schema metadata, entity relationships, and constraints.
Going beyond static documentation, Synapse features a live Data Quality Engine that profiles actual database health (completeness, null anomalies) on the fly.
The AI Copilot: The core of Synapse is powered by the Google Cloud Gemini 2.5 API. Instead of reading technical wikis, users can chat directly with their data architecture. The AI understands the extracted topology, auto-generates business-friendly table summaries, explains foreign-key relationships in plain English, and can even write production-ready SQL joins. Finally, users can export this contextualized metadata as JSON or Markdown artifacts with a single click.
Tech Stack: * Frontend: Streamlit with custom Glassmorphism CSS UI
-
Backend: Python, Pandas, SQLite
-
AI Engine: Google Cloud Gemini API
-
Dataset Validated On: Olist Brazilian E-Commerce (~100k rows)
This build was uploaded as a hackathon project
