Development of dual camera dashboard cams with embedded neural networks for drowsiness/sleepiness detection and road objects detection. Created 7 prototype generations and planned testing with national transport companies. Raised over 20,000$.Neural networks of Jolda were trained using the Python-based library PyTorch and the architecture YOLOv8 by Ultralytics. There were two datasets (one has a MIT license) consisting of 23000 images to train a model for the detection of all road objects, including all road signs, cars, and pedestrians. The model was trained on 70 epochs and obtained the following results: Precision: 86%; mAP50: 82%; mAP50-95: 77%.

I’ve started working as a Research Assistant with Professor Yi Fang in the Laboratory of Artificial Intelligence and Robotics (AIR) at NYUAD. Currently, I'm contributing to the development of the humanoid robot Unitree H1.