ExamShield AI Rover integrates edge-deployed computer vision, autonomous rover surveillance, and centralized exam management on AWS to detect unauthorized devices and verify student identity during high-stakes assessments. A Raspberry Pi 4 with MTCNN face detection and YOLO device recognition runs locally, feeding identity verification and alert data to a backend management system backed by Supabase. An ESP32-controlled rover with ultrasonic sensors and servo motors provides physical surveillance, while the Management Dashboard gives proctors real-time visibility into exam integrity. This architecture combines edge AI inference for low-latency detection with cloud-based alert aggregation and audit trails, eliminating network dependency during critical exam moments. Fork this diagram on Diagrams.so to customize sensor configurations, add additional rover behaviors, or integrate alternative face detection models. The design demonstrates hybrid edge-cloud ML deployment patterns ideal for regulated testing environments requiring both autonomy and centralized oversight.