AI Sleep Monitoring and Enhancement System

An AI sleep analysis system built with sleep scientists. A deep CNN classifies sleep stages from smartwatch sensor data with 94.2% accuracy, using under 18% of the watch battery overnight, and cuts average sleep latency by 22 minutes. Predicts REM and disrupted sleep, gives personalised guidance, and includes a clinical-grade dashboard for healthcare professionals. Delivered in 14 weeks by a 5-person team and has completed human trials ahead of launch.