Video Analysis & Management System for School Surveillance

Designed and developed a distributed real-time surveillance system with AI-based video

analysis on edge devices (Jetson Nano), reducing response time to safety incidents by 85%

o Built a scalable backend using Node.js and BullMQ with Redis for job scheduling, processing

500+ clips/day and cutting storage costs by 40% through TTL-based cleanup

o Architected a hybrid storage model (local + cloud metadata via MongoDB) with retention

policies managing 2TB+ of data across 50+ devices