Model Quality Evaluator - English
Outlier (by Scale AI) London US$15 ph
Review AI-generated media by roleplaying live conversations with voice models and scoring each exchange, comparing image and video pairs against prompts, and labelling first-person task videos. You inspect outputs closely for artifacts like distorted anatomy or mismatched audio, then write specific evidence-based explanations. Work is fully remote and self-scheduled on a freelance basis.