Data Labeller (Atlas, Remote)
Reviewed and refined AI-generated video annotations in line with detailed project guidelines for machine learning datasets. Ensured correct timestamp alignment and accurate action labels and object interactions across high-volume tasks. Verified consistency and quality of annotations while resolving ambiguities with remote project teams. • Corrected timestamp alignment for video segments • Refined action labels and object interaction annotations • Checked guideline compliance and dataset consistency • Communicated with teams to improve labeling quality