Smart India Hackathon 2026 · Problem Statement SIH26192

Minutes that turn panic into preparedness.

PRAVAAH (प्रवाह) is an AI-driven, IoT-fused early warning and evacuation decision platform engineered specifically for the fast-flowing mountain catchments of India's hilly terrains.

The Ground Reality

Why generic weather forecasts fail in Indian hill villages

In the plains, floodwaters rise gradually over days, giving authorities ample time to broadcast evacuation orders. But in the steep gorges of Uttarakhand, Himachal Pradesh, Sikkim, and the Western Ghats, flash floods triggered by cloudbursts or glacial breaches unleash destructive torrents in just 15 to 30 minutes.

Traditional meteorological reports issue district-wide alerts covering thousands of square kilometers. By the time a village sarpanch or citizen realizes that their specific tributary has swelled beyond its banks, the evacuation route is already severed.

Pravaah bridges this life-or-death gap. We don't just predict rainfall; we fuse upstream river sensor telemetry, digital elevation models (DEM), and hydrological velocity modeling to answer the single question a village needs: “How many minutes do we have, and which shelter is safe to reach?”

87,474High Vulnerability

Active landslide and flash flood zones across 179 districts in 19 Indian states.

Source: Geological Survey of India (GSI)

12.6%Terrain Exposure

Of India's land area is susceptible to rapid slope runoff and cloudburst inundations.

Source: NDMA & National Disaster Records

15 minTarget Lead Time

Pravaah calculates dynamic lead time to allow orderly evacuation to concrete shelters before river crest arrival.

Calibrated for rural hill walking speeds

Meet Team Binary Bandits

Hrishiraj Chowdhury

Hrishiraj Chowdhury

ML Engineer

I develop robust machine learning models and intelligent data pipelines to power AI solutions.

Key Contribution:

Developed the core runoff prediction models, hydrograph simulations, and intelligent data pipelines powering Pravaah's early warning engine.

Ashutosh Sharma

Ashutosh Sharma

Full Stack Dev

I bridge the gap between server and client sides, delivering scalable and fully integrated web solutions.

Key Contribution:

Engineered full-stack APIs, backend data models, real-time alert integration, and resilient client-server synchronization across Pravaah.

Krishna M Singh

Krishna M Singh

AI Engineer

I specialize in deep learning algorithms and building advanced artificial intelligence architectures.

Key Contribution:

Specialized in deep learning architectures and predictive algorithms for flash flood crest propagation and temporal risk classification.

Akshita Manker

Akshita Manker

UI/UX Designer

I create intuitive and aesthetically pleasing interfaces that ensure a seamless user experience.

Key Contribution:

Crafted the accessible emergency visual design system, high-contrast disaster alerts, and user experience tailored for low-bandwidth village access.

Prasun Kumar

Prasun Kumar

Frontend Lead

I lead the frontend architecture, focusing on highly responsive and performant web applications.

Key Contribution:

Architected the Next.js application, interactive risk mapping interfaces, real-time alert toast notifications, and client state orchestration.

Poorva Srivastava

Poorva Srivastava

Data Engineer

I design and maintain large-scale data processing systems to handle complex data workflows.

Key Contribution:

Designed and maintained high-throughput data processing systems handling real-time IoT river gauge telemetry and weather data streams.

Visit Github Repository for Source Code

Visit Repository