UBER AI - Data Annotator
In my role at Uber AI, I used the Data Compute platform to label and evaluate video data by comparing generated outputs against original input videos. I assessed key attributes such as motion consistency, ensuring movements were natural and aligned with the source material, as well as identifying any discrepancies in visual continuity. I also evaluated sensitive attributes such as race and gender representation, carefully checking for unintended changes or inconsistencies between the input and output videos. This required a high level of attention to detail, strong analytical skills, and strict adherence to evaluation guidelines. My work contributed to improving the fairness, accuracy, and overall quality of AI-generated video outputs.