This project involved creating a surveillance system with object detection and tracking deployed on an edge device. The system was designed to process 19 RTSP video streams simultaneously at 8-10 FPS each. Leveraging YOLOv8 for object detection and NvDCF for tracking, the solution was built on Nvidia Jetson AGX Xavier hardware. The goal was to optimize performance while maintaining high resolution (2K) and real-time processing without relying on costly cloud computing solutions.