AI Training and Data Labeling
Automotive AI – Object Detection and Segmentation: Labeled and validated image and video datasets for vehicle perception systems. Tasks included bounding box and polygon annotations for objects such as cars, pedestrians, and traffic signs. Collaborated with model engineers to analyze labeling errors and improve data consistency for deep learning model training. Synthetic vs. Real Data Comparison: Supported model training using both synthetic and real-world videos to improve domain adaptation and robustness. Helped generate synthetic datasets to enhance model generalization and reduce dependency on real data.