// Service
Tarek Abou Rjeily builds real-time computer-vision systems for clients in Lebanon and the Gulf — software that sees what people can't, then detects, matches, and alerts the instant it happens. Examples include a real-time exam-proctoring prototype (YOLOv8 + face recognition) and a face-recognition photo-delivery platform for events.
A real-time exam-integrity prototype. YOLOv8 + face recognition flag phones, head-turns, and unknown faces; alerts include the flagged frame in Telegram.
Read case study →Face-recognition photo delivery for events. One selfie returns matching event photos — built on a self-hosted face engine, with guest consent and per-event isolation.
Read case study →Tarek Abou Rjeily builds real-time exam proctoring, face recognition and matching, behaviour and object detection, and visual quality checks, using YOLOv8, OpenCV, MediaPipe, and self-hosted face engines like CompreFace.
Privacy has to be built into the product. In BSoura, for example, guests see a consent screen before their selfie is matched, each event is isolated behind its own private link, and organisers choose whether selfies are kept, so a face is only ever matched inside one event.
Message me on WhatsApp, email tar@tarekrjeily.com, or see all services.