Data Annotator
TextRLHF
I contribute to Reinforcement Learning from Human Feedback (RLHF) by serving as a human rater for AI model training and alignment. I compare and rank responses from competing model variants across multi-turn prompts, evaluating them for correctness, helpfulness, and quality using structured rubrics. My evaluations span coding (Python, PowerShell, Vue/Quasar, pandas), mathematics, creative writing, and SRE/DevOps tasks, providing the preference signals that guide model improvement. I also identify edge cases, flag problematic outputs, and ensure consistent application of rating criteria to maintain high data quality across evaluation workflows.
I contribute to Reinforcement Learning from Human Feedback (RLHF) by serving as a human rater for AI model training and alignment. I compare and rank responses from competing model variants across multi-turn prompts, evaluating them for correctness, helpfulness, and quality using structured rubrics. My evaluations span coding (Python, PowerShell, Vue/Quasar, pandas), mathematics, creative writing, and SRE/DevOps tasks, providing the preference signals that guide model improvement. I also identify edge cases, flag problematic outputs, and ensure consistent application of rating criteria to maintain high data quality across evaluation workflows.
2025 - Present