AI Trainer / Data Annotation Specialist (Mercor AI) — AI-generated text evaluation and annotation for LLM alignment
Evaluated AI-generated responses for accuracy, relevance, coherence, factual correctness, and safety using project annotation guidelines. Performed comparative response ranking and preference evaluation to support RLHF and model alignment workflows. Identified hallucinations, logical inconsistencies, bias issues, and factual inaccuracies while maintaining high quality and consistency standards. • LLM response quality assessment and rating • Hallucination and factuality detection • Safety and compliance-oriented evaluation • Comparative preference/ranking for RLHF