Data Labeling Specialist (Atlas Capture)
Performed high-precision image annotation for computer vision training datasets, including bounding boxes, polylines, and semantic segmentation. Followed strict annotation guidelines to ensure label consistency and achieve 98% inter-annotator agreement (IAA). Completed pre-final dataset quality assurance to identify and correct discrepancies before delivery. • Labeled dataset items using bounding boxes and polyline-based annotations • Applied semantic segmentation for visual classes • QA on 10,000+ images, verifying label correctness • Fed back improvements to the Atlas Capture engineering team