Mohamed Hannat

Advancing Industry 4.0 Automated Inspection Solutions:

○ Lead the development of multimodal vision systems (CLIP, SAM) achieving 97% detection accuracy

and 35% reduction in processing time

○ Spearhead IoT projects integrating edge AI with NVIDIA Jetson platforms using TensorRT acceler

ation (35+ FPS)

○ Designed and implemented network of PoE cameras for high-quality data capture in industrial

environments

○ Designed and implemented MLOps pipeline with CI/CD for model versioning and A/B deployment

strategies

○ Integrated AWS services (SageMaker, Lambda, API Gateway) for scalable computer vision infras tructure

○ Deployed AWS IoT Core to manage device fleets with secure MQTT communication for real-time telemetry

○ Delivered 4 major IoT projects: Almarai, Saudi CAN, PSI, and KACST, generating estimated 2.3M

annual cost savings

○ Integrated cross-modal retrieval systems for intelligent industrial data analysis

Trajectory Planning for Robot Navigation:

○ Developed trajectory planning software for Doosan robots, enabling efficient and accurate au

tonomous navigation

○ Fine-tuned HuggingFace BLIP-2 model for vision-language understanding in dynamic industrial

environments

○ Optimized computer vision pipeline with TensorRT for real-time performance (30+ FPS)

○ Implemented AWS-based infrastructure using Lambda and DynamoDB for robot telemetry storage

and analysis

○ Collaborated with mechanical and software teams to ensure seamless integration and operation of the navigation system

Provided hands-on coaching in robotics and AI, guiding learners in building smart systems using Python, machine learning, and embedded technologies. Designed and delivered practical workshops and project-based training sessions.