Hasan Öğüt

For my bachelor's thesis, I built a real-time pedestrian detection and time-to-collision warning system on NVIDIA Jetson Orin Nano. Its zero-copy memory architecture removes the CPU-GPU bottleneck, and my FP16/FP32 optimization of eight object detection models achieved up to 157% higher throughput.

At One Eyes System, where I was an intern, I developed real-time person detection, pose estimation, and keypoint extraction pipelines on edge devices. I also optimized inference latency and frame rate on resource-constrained hardware, contributing to model selection and deployment.

With Yılkı TEAM, I served as Team Captain on an AI-based decision support system for real-time disaster coordination, which placed 12th in Türkiye at TEKNOFEST 2025. I'm also developing a Unity simulation environment for design validation, virtual commissioning, and Vision AI synthetic data generation for robotics and industrial machinery.