AI-Driven CBT Proctoring Setup
Publicada el 2026-07-22
Descripción de la oferta
I need my existing computer-based testing platform re-configured so that AI handles proctoring and live monitoring end-to-end. The goal is for the system to watch the screen, webcam, and microphone feeds concurrently, flag anything out of the ordinary, and respond on the spot. Core requirements • Continuous capture of screen activity, webcam video, and ambient audio during each session. • On-device or server-side AI that detects potential cheating behaviours in real time. • Immediate actions when a rule is broken: – push a real-time notification to an invigilator dashboard or Slack channel, – start/append a detailed log with time-stamped recordings, – terminate the test automatically if the violation crosses a preset threshold. Please wire the solution into the current CBT code-base (PHP/Laravel on the back end, React on the front if you need specifics), keeping latency low and storage use reasonable. Any third-party libraries—OpenCV, TensorFlow, or proprietary SDKs—are fine as long as licences allow commercial use. Deliverables 1. Fully integrated AI proctoring module running in our staging environment. 2. Deployment instructions and config files so the team can promote to production. 3. A short README that explains how alerts are generated, where recordings are stored, and how thresholds can be tuned. Acceptance criteria • A demo session shows all three data streams captured and analysed live. • Triggering a test violation fires the three required actions without manual refresh. • No existing test functionality is broken after the integration.
Skills
Fuente original: freelancer