Freelance AI Data Labeler & Domain Expert — RLHF Ranking, Evaluation & Prompting (2023–Present)
Performed RLHF-style ranking and rating of AI-generated outputs using detailed rubrics. Compared multiple model responses and selected the best answer based on factual correctness, logic, helpfulness, and safety. Conducted expert evaluation and iterative feedback to improve model reasoning and output quality. • Ranked and rated responses by quality and accuracy for RLHF pipelines • Evaluated coherence, correctness, and safety against rubrics • Provided structured expert feedback to guide model improvement • Supported prompt authoring and iterative refinement for coverage and diagnostic value