SevaSetu
AI Disaster Response Platform
A zero-trust, AI-verified disaster response platform that turns distress reports into verified, proximity-based relief actions.
React 19
Node.js
Express
SevaSetu is an AI-driven humanitarian logistics and disaster response platform focused on eliminating unverified reports and coordination delays. It validates incoming distress data, prioritizes urgency, and routes missions to trusted responders.
How it's built
Frontend is built with React 19 + Vite. A Node.js/Express backend handles geospatial orchestration with PostgreSQL/PostGIS, while a dedicated FastAPI service runs vision and metadata verification pipelines.
What it does
- AI report validation with CLIP + EXIF integrity checks
- Radius-based volunteer broadcast dispatch
- Coordinator command dashboard with urgency scoring
- Offline capture and deferred sync for field resilience
Where it got hard
Balancing high-frequency real-time updates with backend scalability while keeping verification strict enough to prevent fraudulent or stale reports.
This project strengthened my ability to combine geospatial systems with AI validation in safety-critical workflows where trust and response time are equally important.
What's next
Expand multilingual support, improve dispatch optimization heuristics, and deepen integration across messaging channels for faster citizen reporting.