Back to Projects

SevaSetu

AI Disaster Response Platform

A zero-trust, AI-verified disaster response platform that turns distress reports into verified, proximity-based relief actions.

SevaSetu | AI Disaster Response Platform

Technologies

React 19ViteNode.jsExpressPostGISFastAPIGeminiLeaflet

Overview

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.

Architecture & Decisions

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.

Key Features

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

Challenges

Balancing high-frequency real-time updates with backend scalability while keeping verification strict enough to prevent fraudulent or stale reports.

Lessons Learned

This project strengthened my ability to combine geospatial systems with AI validation in safety-critical workflows where trust and response time are equally important.

Future Roadmap

Expand multilingual support, improve dispatch optimization heuristics, and deepen integration across messaging channels for faster citizen reporting.