Building NIRNAY: How We Used AWS Amplify & Amazon Bedrock to Solve Delhi's Urban Waterlogging Crisis

Discover how NIRNAY combines Euler numerical hydrology, dynamic programming, and Amazon Bedrock Claude 3.5 Sonnet to shift Delhi's monsoon flood response from reactive firefighting to proactive, sub-second decision making.

Vikesh

Published Oct 11, 2026

Last modified Oct 11, 2026

🌊 The Urban Crisis: When Delhi Paralyzes in 30 MinutesEvery monsoon season, the capital city of India experiences an alarming recurring crisis. Intense cloudbursts and torrential downpours transform critical traffic arteries and low-elevation underpasses—such as Minto Bridge, Zakhira, Tilak Bridge, and Pul Prahladpur—into submerged urban reservoirs within 30 to 45 minutes.
The consequences are catastrophic:
Emergency ambulances heading to premier trauma centers like LNJP Hospital and AIIMS get trapped in gridlocks.
Public transit buses submerge up to their windshields.
Civic bodies like the Municipal Corporation of Delhi (MCD) and Public Works Department (PWD) are forced into reactive firefighting—deploying emergency mobile diesel suction pumps only after roads are already underwater and traffic is stranded.
As engineering students and builders, our team asked a fundamental question:
What if civic authorities could predict exact flood depths and mathematically optimize pump deployment before the storm arrives?
This inspired us to build NIRNAY (National Infrastructure Risk & Network Action Yield)—an AI-powered Pre-Storm Decision Support Engine.
﻿
🏗️ System Architecture & How We Built It on AWSNIRNAY is engineered around a high-performance full-stack architecture that combines deterministic physics modeling with generative AI situational awareness, powered end-to-end by AWS.
text
┌────────────────────────────────────────────────────────┐
│               Client / Field Operations               │
│         Official MCD Civic Control Room UI            │
│         (React 18, TypeScript, Tailwind CSS)          │
└───────────────────────────┬────────────────────────────┘
                            │
              Hosted on AWS Amplify Hosting (CI/CD)
                            │
┌───────────────────────────▼────────────────────────────┐
│              FastAPI & Hydrology Engine                │
│  - Copernicus 30m DEM Terrain Elevation Integration    │
│  - Rational Runoff & Euler Numerical Inflow Model      │
│  - Dynamic Programming (DP) Knapsack Pump Optimizer    │
└───────────────────────────┬────────────────────────────┘
                            │
       AWS Boto3 SDK / Agentic Tool Orchestration
                            │
┌───────────────────────────▼────────────────────────────┐
│                   Amazon Bedrock                       │
│    Claude 3.5 Sonnet Grounded Conversational Agent     │
│   (Translates physics math into bilingual advisories)  │
└────────────────────────────────────────────────────────┘
﻿

# agents # aws-amplify # aws-app-runner # python # aws-dms-serverless

Any opinions in this article are those of the individual author and may not reflect the opinions of AWS.

Enjoyed reading this content? Let the author know!

Your likes, comments, shares, and saves help creators reach more builders.

Loading recommendations

Loading article

Builder Center

Building NIRNAY: How We Used AWS Amplify & Amazon Bedrock to Solve Delhi's Urban Waterlogging Crisis

Vikesh

🌊 The Urban Crisis: When Delhi Paralyzes in 30 Minutes

🏗️ System Architecture & How We Built It on AWS