DeliriumWatch

Real-time monitoring pipeline built for Grand River Hospital staff, with secure login and role-based access control. The system was designed to reduce manual environmental monitoring during simulated testing — we saw a 90%+ reduction in that workload.

I implemented OpenCV-based eye detection and a live Flask-powered web visualization with automatic alert flagging, enabling automated sleep and blink classification and faster response to abnormal conditions during validation. On the hardware side, I engineered serial data ingestion and threshold validation for stable multi-sensor tracking with zero dropped readings.