AI Training Specialist / Domain Expert (Mercor Contractor)
As a Domain Expert contractor with Mercor, I evaluated and ranked over 200 pairs of AI-generated text responses using client-supplied rubrics. My work supported RLHF fine-tuning pipelines by providing preference judgments and written rationales. Attention was paid to identifying hallucinations, age-inappropriate content, and curriculum misalignment in educational materials.• Reviewed and ranked AI response pairs for accuracy and alignment with rubric criteria. • Supported reinforcement learning with human feedback (RLHF) by providing structured preference data. • Maintained a 95% acceptance rate across all contracts with no revision requests. • Flagged potential issues in AI output such as hallucinations and inappropriate lesson content.