Designed and built a full-featured network video recorder (NVR) system using Django to replace cloud-dependent security solutions with a private, locally hosted alternative running on a home server.
Integrated Power over Ethernet (PoE) cameras and processed RTSP video streams with OpenCV, enabling live viewing and real-time frame capture directly within the Django application.
Implemented AI-powered object detection using YOLOv8n, drawing bounding boxes around detected objects in live streams — allowing motion-aware monitoring without cloud dependency.
Built a custom calendar-based interface for browsing and retrieving recorded footage, making it easy to find specific events across large libraries of video data.