NadirPoise: AI-Powered Safety Shift Scheduling

Night shifts carry a dangerous blindspot that costs lives and billions of dollars. Studies from the National Safety Council show workplace accidents spike by 30% between 2:00 AM and 5:00 AM—a biological state known as the "Nadir Window" when human alertness hits its lowest point. Fatigue costs global enterprise operations over $130 billion every year in workplace injuries, equipment damage, and legal liabilities. Yet, legacy scheduling software still treats humans like static machines, completely ignoring mandatory safety frameworks under ICAO and EASA regulations. NadirPoise solves this problem by transforming basic shift creation into an intelligent, biologically grounded, and fully auditable safety workflow. (Built entirely in native.builder for the AI Factory hackathon.) How NadirPoise Works in 5 Clear Steps: 1. Physics-Based Calibration (NASA POWER API): Pulls real-time geographic solar irradiance data based on exact workplace coordinates to base fatigue calculations on physical facts rather than AI assumptions. 2. 5-Persona AI Safety Test (AIML API): Before a roster is published, our AI engine (running Llama 3.3 70B) tests the schedule from 5 operational viewpoints: Chief Fatigue Officer, Shift Supervisor, Night Ops Lead, HR/Wellbeing Director, and Compliance Auditor. 3. Automatic Lock & Voice Verification (Speechmatics API): If a schedule exceeds fatigue safety limits, the system automatically locks execution. A manager must provide a clear written or spoken justification to unlock it. 4. Digital Compliance Receipts: The AI evaluates the manager’s justification against safety standards and generates a tamper-proof receipt ID (e.g., ICAO-EASA-AIML-VERIFIED-XXXX) ready for official audits. 5. Worker Recovery App: Frontline workers receive simple, personalized guidance on rest and light exposure to help them recover quickly from sleep debt. Built end-to-end on native.builder with Supabase managing real-time data sync.