Description:
The Carbon Footprint Tracker is an AI-powered platform designed to help industries monitor, analyze, and reduce their carbon emissions in real time. It uses IoT sensors to collect environmental data and machine learning models to perform multistep prediction of carbon output. The system provides actionable insights based on the Air Action Plan (AAP), Emissions Trading System (ETS), and transport-related emissions, enabling organizations—especially red-zone industries—to proactively reduce their environmental impact.

Key Features:

• 📡 Real-time monitoring using IoT devices
• 📊 Multistep emission prediction using CNN-LSTM
• 🧠 AI-driven insights for emission reduction
• 🔍 Industry-specific recommendations
• 📈 Historical trend analysis
• 🌐 Web dashboard for decision-making

Tech Stack: Python, TensorFlow, CNN-LSTM, CI/CD, Flask, React.js, Node.js, Supabase, IoT Sensors

Description:
A real-time AI-powered Network Intrusion Detection System that monitors and analyzes network traffic for suspicious activities. Using anomaly-based techniques, it detects deviations from normal behavior to identify zero-day attacks and stealthy intrusions.

Key Features:

• 🚨 Anomaly-based intrusion detection
• 🧠 Intelligent pattern recognition for zero-day attacks
• 📈 Real-time traffic monitoring
• 🧩 Visualization dashboard for threat reports
• 🔒 Modular design for IPS integration
• 🔁 Scalable and adaptable to evolving networks

Tech Stack: Python, Scikit-learn, XGBoost, Random Forest, Flask, Wireshark, Pandas, Matplotlib

Description:
A real-time AI-driven safety solution that predicts threats on campus before they occur. Instead of detecting objects, it analyzes posture, gestures, and facial expressions to understand intent. The system identifies risks like assaults, intrusions, theft, and fires, then triggers automated IoT responses and emergency alerts.

Key Features:

• 🔍 Intent-based threat prediction using ResNet-50 + Bi-LSTM
• 🗺️ Risk hotspot mapping
• 📡 Human-in-the-loop alerts via SMS and calls
• 🔔 IoT auto-response mechanisms
• ⚡ Real-time monitoring dashboard
• 🔑 Reduced false alarms through intent analysis
• 🏫 Full campus safety integration

IoT Auto-Actions:

• Intruder → doors lock, lights off, siren on
• Fire/smoke → doors open, lights on, evacuation siren

Tech Stack:
AI: ResNet-50, Bi-LSTM, Custom Dataset
Hardware: ESP32, MQ-135, SG90 Servo, LEDs, Speaker
Frontend: React, Tailwind
Backend: Supabase

Description:
An advanced AI-powered legal research engine tailored for commercial courts. It extracts, analyzes, and recommends relevant judgments using NLP and machine learning models, improving the speed and accuracy of legal research.

Key Features:

• 🧠 NLP-based legal text processing (LegalBERT, SpaCy, T5)
• 🔍 Semantic search & summarization
• 📚 Predictive analytics for case outcomes
• 🌐 Multilingual support (XLM-R)
• ✅ Personalized case suggestions
• 📊 Interactive dashboard with filters & highlights

Tech Stack: Python, LegalBERT, T5, XLM-R, SpaCy, Elasticsearch, Scrapy, Flask, Django, React.js, SHAP, XGBoost, LSTM

|

• LinkedIn: Het Mehta
• Portfolio: hetmehtaportfolio.vercel.app