Award Winning AI Research Project

AI-Powered Disaster
Intelligence for India

Predicts disaster risks and utilizes an advanced uncertainty gating mechanism to determine if those predictions are reliable enough for action. Reducing unsafe operational decisions by up to 46%.

Megavarshan A

Megavarshan A

Lead Developer & Researcher

🎓 Final Year AI Undergrad at SRMIST, Chennai

🚀 Co-Founder, Foresight-X Research Labs

🏆 Award Winner at SRM Research Day 2026

46%

Unsafe Decision Reduction

Nominal Conditions

35%

Unsafe Decision Reduction

Anomalous Data

69.2%

Coverage Rate

Events Processed

12.98%

Gated Unsafe Rate

vs 24.19% Baseline

The Core Innovation

Instead of only predicting disasters, the system assesses uncertainty to predict if its own prediction should be trusted.

Risk Assessment

Multi-hazard predictions

Reliability Estimation

Prediction confidence

Anomaly Detection

Data shift metrics

Smart Gating: If confidence is high and data is normal, the alert executes. Otherwise, the system abstains — preventing false alarms.

Platform Capabilities

Integrated modules for comprehensive Indian disaster intelligence

Live Monitoring

Real-time tracking of cyclones, tsunamis, floods, and earthquakes.

Geospatial Intelligence

Interactive maps with hazard markers, risk zones, and NDRF deployment locations.

Reliability Engine

Evaluates the confidence and uncertainty of AI predictions in real-time.

Smart Gating

Core innovation — AI that knows when NOT to decide based on data anomalies.

Multi-Channel Alerts

Send real-time warnings via SMS, Telegram, and WhatsApp to affected populations.

Public Reporting

Citizens can report incidents directly, reflected in government dashboards.

Multilingual AI

Voice and text assistant in English, Tamil, Hindi, Kannada, and Telugu.

India-Focused

Designed for Indian disaster management with IMD, INCOIS, CWC, NDRF integration.

Award-Winning Research

Built on state-of-the-art AI reliability frameworks

Megavarshan A

Lead Researcher • SRMIST

Developed an advanced disaster intelligence platform utilizing a novel reliability-aware architecture. The system intelligently integrates real-time hazard prediction with uncertainty estimation and data shift monitoring to dynamically gate critical operational decisions.

Learn more about the developer