Doruk Sönmez, M.Sc., is an AI Solutions Architect at CTai LABS, Connect Tech Inc.’s dedicated Physical AI and Edge AI engineering department. He holds a Master’s degree in Computer Engineering with a concentration in Deep Learning from Gazi University (2021 to 2023). His thesis focused on an end-to-end optimized artificial neural network deployed on NVIDIA Jetson platforms, redeveloped for autonomous vehicles and benchmarked against other state-of-the-art models, work that led directly to a peer-reviewed paper.

Before joining Connect Tech, Doruk spent several years at OpenZeka in Turkey as an Intelligent Video Analytics Engineer, developing computer vision applications for AI and robotics projects on embedded platforms. He presented an Early Fire Detection application and a Civil and Worker Security Equipment Detection application at GITEX Dubai 2019, spoke at an NVIDIA Jetson meetup in Ankara on AI application demos across DeepStream, TensorRT, and YOLO, and led training on optimized neural network development at OpenZeka’s MARC competition. He also maintained a multi-year, bylined presence on OpenZeka’s Turkish-language technology blog, writing and translating articles on topics including transformer models and NVIDIA Jetson-powered autonomous agriculture. At GTC 2024, OpenZeka’s own channels featured him demonstrating the Cordatus AI platform on a Jetson Orin device.

His academic work culminated in a peer-reviewed paper, Sönmez, D. and Çetin, A. (2024), “An End-to-End Deployment Workflow for AI Enabled Agriculture Applications at the Edge,” presented at the 2024 6th International Conference on Computing and Informatics (ICCI) in Cairo, Egypt, and co-authored with his Gazi University advisor. The paper has since been cited by other published research in the precision agriculture and edge AI space.

At CTai LABS, Doruk focuses on deploying Physical AI directly at the Edge, connecting simulation, vision analytics, and real-world sensor data into a single production workflow. At NVIDIA GTC 2026 in San Jose, he was quoted on the record describing CTai LABS’ live demonstration of a Real2Sim pipeline, built on NVIDIA’s Video Search and Summarization Blueprint and NVIDIA Cosmos Reason2, running on the Connect Tech Anvil-T5 Edge System powered by NVIDIA Jetson Thor.