Ahmed Hermas
Senior AI Engineer
Seven years colaborating and leading teams, building real-world computer-vision systems, from ISO-certified biometrics to city-scale traffic AI to Agentic AI to 3D vehicle reconstruction. Below are the projects I shipped.
Biometri — ISO-Compliant E-KYC
Led R&D and delivery of the face-recognition and anti-spoofing stack.
A facial recognition platform engineered for high-stakes identity verification. I directed the research, training, and productization of the core face engine and the presentation-attack detection layer, successfully earning the system its ISO/IEC 30107 certification. This solution was seamlessly deployed into a diplomatic payment gateway and a regional educational attendance system running efficiently on Edge hardware.
Vision AI Vehicle Inspection Center
From Faulty prototype to production-ready solution.
A computer-vision queue-management product designed for high-volume vehicle-inspection centers, where lane throughput directly dictates operational KPIs. I owned the hardening phase—transforming an initial pilot into a robust, production-ready system capable of withstanding extreme field conditions. The final architecture runs end-to-end on a single embedded edge device, eliminating cloud latency and dependencies.
Smart Cities — Traffic Intelligence Platform
Directed model optimization and edge deployment of an 8-model AI suite.
A city-scale traffic-intelligence platform powered by eight cooperating computer-vision models, enabling multi-camera tracking, sensor-fusion speed estimation, in-cabin behavior detection, and an algorithmic reckless-driving classifier. I led the migration of the entire software suite from server-grade GPUs to NVIDIA Jetson AGX edge units in production. This effort maintained strict accuracy thresholds while unlocking real-time, multi-camera processing capabilities at the edge.
Cross-Camera Journey Analytics Platform
Lead Architect: End-to-end research, pipeline development, and deployment.
An end-to-end multi-camera tracking system developed for a major national energy provider, designed to stitch together the complex journeys of vehicles and pedestrians across expansive service station footprints. I independently managed this project from the initial research phase—selecting and optimizing the re-identification stack—through to full deployment. The system successfully translates raw camera feeds into actionable operational dashboards, significantly strengthening the security posture for site stakeholders.
3DAI Vehicle Inspection and Audit
Managed the end-to-end proof of concept architecture and delivery.
A pioneering proof-of-concept initiative translating advanced Gaussian splatting algorithm into practical, industrial vehicle-inspection workflows. I led the development of a complete pipeline capable of capturing a vehicle in-bay and reconstructing it in high-fidelity 3D. The resulting operator-facing navigable model provided audit-grade documentation, equipping inspectors with next-generation spatial visualization tools that revolutionized their standard operational procedures.
Let’s build something
Open to interesting opportunities, advisory work, and collaborations. The fastest way to reach me is email.
- Emaila7medhermas@gmail.com
- Phone+971-522902006+20-1112440020
- LocationUAE ✈ Egypt
© 2026 Hermas AI — Built with Next.js, TailwindCSS & vibe coding.