At Atlas, I work on ego-view video annotation to train AI activity recognition models
At Atlas, I work on ego-view video annotation to train AI activity recognition models. I watch first-person footage recorded from the perspective of the person using the device, and produce detailed step-by-step descriptions of every action they perform — for example, picking up a plate, rinsing it under water, placing it on a rack to dry. The goal is to give the AI an accurate, granular understanding of everyday human activities like washing dishes, folding clothes, arranging items, and cleaning. I work within Atlas's internal annotation platform, following structured guidelines to ensure consistency and accuracy. I have maintained a 98–100% accuracy rate throughout my time on the platform.
Not specified