Fortix AI — Intelligent Video Surveillance Platform

Developed a comprehensive AI-powered video surveillance platform that leverages computer vision for real-time object detection, violation monitoring, and automated alert systems. The platform integrates multiple AI models for diverse detection scenarios including trash detection, hardhat safety compliance, fire/smoke detection, and general object recognition.

Key Features

Automated Scheduling & Resource Management: AWS Lambda functions for automated pod start/stop operations
EventBridge Scheduler for time-based GPU resource optimization
Dynamic pod creation/termination based on availability
Real-time Monitoring & Alerts: Live video streaming with overlay annotations
Push notifications via Expo
Email alerts with image attachments (AWS SES)
Advanced Video Analytics: AI-powered video summary generation
Automated report creation with violation timestamps
Export capabilities (PDF reports)

Technical Architecture

Frontend: Deployed on AWS EC2
Backend: Node.js API servers on AWS EC2
AI Models: Dockerized Python FastAPI application deployed on RunPod GPU instances

Highlights

Developed a comprehensive AI-powered surveillance system with real-time object detection, integrating custom-trained YOLO models and GPU-based inference.