Egemen Kopuz

Thesis Title: Multi-Modal 3D Object Detection in Long Range and Low-Resolution Conditions of Sensors - Conducted Master’s Thesis on multi-modal (LiDAR and Camera) 3D object detection for intelligent transportation systems and autonomous railway domains. - Developed a fusion deep learning model in PyTorch with a pipeline integrating temporal data which enhanced detection accuracy significantly, especially for distant and occluded objects, and low-resolution sensors.

- Researched deep learning and machine learning models for quality evaluation through acoustic/audio classification in varying automotive use-cases, utilizing PyTorch and Scikit-learn for implementation. - Designed and implemented an internal website for model inference and analysis using Flask framework for back-end, Bootstrap and jQuery for front-end, and SQLite for the database. - Deployed and maintained production code on AWS EC2 using Docker.