AI Trainer / Aether Generalist - Outlier
As an AI Trainer and Aether Generalist at Outlier, I evaluated AI-generated responses for correctness, clarity, and instruction adherence using structured workflows. I annotated and labeled both text and image datasets to support AI model training and improved response selection through quality comparison. Logical inconsistencies, factual inaccuracies, hallucinations, and reasoning steps were consistently identified and reviewed for coherence. • Evaluated text responses within LLM evaluation workflows. • Performed data annotation and labeling to improve model training on text and images. • Selected higher-quality AI responses by comparing multiple outputs based on reasoning and factual accuracy. • Reviewed and flagged logical, factual, and coherence issues in AI outputs for model feedback.