Remotask
Worked on a geospatial data labeling project focused on autonomous driving use cases, involving annotation of tiled imagery and LiDAR-derived visual data within the Remotasks platform. The project required accurately identifying and labeling road elements, vehicles, pedestrians, and environmental features using bounding boxes, polygons, and key points to support object detection and scene understanding models. Responsibilities included drawing precise annotations in accordance with detailed labeling guidelines, handling edge cases such as partial occlusion and overlapping objects, and ensuring spatial accuracy across large image sets. Quality standards were strictly enforced through internal validation rules and reviewer feedback cycles, requiring consistent adherence to annotation accuracy, completeness, and class definitions. The work contributed to training datasets used for perception and mapping models in autonomous vehicle systems.