AI Training Specialist & Creative Writer (Contract) | Remote | Jan 2024 – Present
Performed RLHF by ranking AI-generated creative content using multi-dimensional evaluation rubrics. Authored detailed rationales (300+ words) to justify ranking decisions and improve model alignment. Identified and corrected hallucinations and logical inconsistencies to raise output quality and protocol compliance. • Ranked creative text responses for helpfulness, honesty, and harmlessness (HHH) adherence • Wrote 300+ word rationales explaining model ranking outcomes • Curated prompt datasets to test model limits and creative storytelling performance • Flagged hallucinations and logical fallacy issues for correction