Advancing Multi-Modal Perception for LLM-based End-to-End Autonomous Driving.
Designed and implemented a method to integrate traffic light and sign ground truth data from CARLA through python API, which reduced the related driving infractions by 23%.
Developed a perception stack with deep learning-based traffic light and sign recognition model which leads to relative driving performance improvement of 14%.
Estimated accurate evaluation of traffic light detection model by k-fold cross-validation training method and generated a dedicated traffic sign dataset for closed loop simulation system (CARLA) containing 4.3k+ images.
Experience in working with high-performance computing clusters (Accessing virtual machines with SSH, Batch job scripting and Slurm workload management).
Major activities: architecture development (Python), dataset preparation, dataset annotation and augmentation, model training, experiment tracking and metrics evaluation.
Skills: Python, PyTorch, OpenCV, YOLO, RT-DETR, Hugging Face, scikit-learn, pandas, Git, Docker, shell scripting (CLI).
Versatile, analytical, and impact-driven engineering professional with 3+ years of experience in AI/ML and autonomous driving technologies. Possesses a strong expertise in computer vision, deep learning, data engineering and HPC platforms through hands-on projects.
Currently advancing cutting-edge mobility solutions through a master’s program focused on AI integration in automated driving.
Discover other experts with similar qualifications and experience
2025 © FRATCH.IO GmbH. All rights reserved.