Justin Chang

Built end‑to‑end SaaS that runs YOLO‑based human & dog action detection on user RTSP streams

Isolated customer data by auto‑provisioning Fly.io Machines per tenant; Celery watchdog gracefully replaces unhealthy nodes, preventing runaway costs

Added facial recognition & Re‑ID post‑processing, enabling automated per‑person analytics every 3 hrs

Integrated Square subscriptions for billing; full stack in Next.js, Django, Celery, Postgres

Built Clip Saver package for seamless YOLO model integration with RTSP streams, featuring graceful start/stop controls and asynchronous callback system for database and cloud storage integration. Non‑blocking callback system stores images to Cloudinary and rows to Postgres; powers personal home‑security deployment