JalRakshak.AI
Microsoft AI Unlock Hackathon · AI for India Track

About JalRakshak.AI

“Jal Rakshak” — Guardian of Water. An IoT + AI system designed to monitor water quality across India's rivers, reservoirs, and distribution networks in real-time, making clean water intelligence accessible to all.

The Problem

Over 163 million Indians lack access to safe drinking water. Contamination from industrial discharge, agricultural runoff, and aging infrastructure affects millions of rural and urban communities.

Traditional water testing is slow, expensive, and infrequent — lab results take days, and field kits require trained personnel. Communities often discover contamination only after illness spreads.

We need affordable, continuous, automated monitoring that works even in remote areas without reliable internet — and that turns raw sensor data into actionable intelligence.

Our Solution

JalRakshak.AI combines low-cost LoRaWAN IoT sensors with a machine learning model to deliver real-time water quality intelligence at scale. The system:

LoRaWAN + TTN Integration

Long-range, low-power IoT nodes transmit pH, TDS, and temperature data using LoRa radio to The Things Network cloud infrastructure.

AI-Powered Prediction

A Random Forest classifier trained on 3,276 water samples predicts potability in real-time, with safety scoring, cause detection, and trend analysis.

Real-Time Alerts

Instant UI alerts when unsafe water conditions are detected — with detailed cause identification and actionable remediation guidance.

Rural Coverage

LoRa technology enables monitoring across terrain where cellular connectivity is unavailable, extending coverage to remote villages and farms.

Low-Power Nodes

ESP32 + LoRa SX1276 hardware transmits once per minute on battery power, enabling months of unattended operation per deployment.

Built for Bharat

Designed specifically for Indian water bodies, agricultural zones, and municipal supply — addressing real contamination challenges India faces.

How It Works

1

Sensor Node

ESP32 microcontroller reads pH, TDS, and water temperature sensors and encodes them into 6 bytes.

2

LoRaWAN Transmission

LoRa SX1276 radio transmits the payload over long range (up to 10km) to a TTN gateway.

3

TTN Uplink

The Things Network decodes the payload using a JavaScript formatter and forwards it to our webhook.

4

AI Analysis

Our Random Forest model predicts potability; threshold-based analysis generates safety scores, cause detection, and trend forecasts.

5

Dashboard Display

The Next.js dashboard polls the API every 60 seconds, rendering live device cards with predictions and charts.

Technology Stack

Next.js 16
TypeScript
LoRaWAN
TTN (The Things Network)
ESP32 + SX1276
Random Forest (scikit-learn)
Recharts
Tailwind CSS
FastAPI (Python model server)
Vercel

Team

Dual Core
Microsoft AI Unlock Hackathon 2026
RR
Raj Rabidas
Leader

3rd Year B.Tech

Metallurgical & Materials Engineering

IIT Roorkee

MR
Mansi Rajput
Member

2nd Year B.Tech

Mechanical & Industrial Engineering

IIT Roorkee