AI Trainer / Data Annotation Specialist (Aether Project) at Outlier AI
Performed AI-generated response evaluation and labeling to improve large language model performance on the Aether project. Applied project guidelines to assess outputs for accuracy, relevance, clarity, tone, and policy compliance. Conducted comparative ranking and quality scoring to produce training-ready judgments for the model. • Annotated responses according to evaluation rubric • Ranked multiple candidate responses by quality • Scored outputs for safety and policy adherence • Flagged hallucinations and factual inconsistencies