Video Annotation Specialist
As a Video Annotation Specialist at AtlasCapture, I focused on annotating ego-centric video datasets, emphasizing hand-object interactions. I applied both atomic (dense) and goal-oriented (coarse) labeling to ensure each video segment was accurately marked. My work contributed to generating high-quality training data for machine learning models, strictly following detailed rules for auditing and segment integrity. • Segmented video data around "reach-to-release" timestamps and intent boundaries. • Maintained a 95%+ accuracy rate during audit cycles by adhering to labeling guidelines. • Synthesized raw video data into structured, sub-20-second segments with distinct goals. • Assisted in troubleshooting edge labeling cases with coarse granularity strategies.