AI Training Expert
As an AI Training Expert at Scale Labs, I led reinforcement learning from human feedback (RLHF) tasks for large language models. I established annotation guidelines and scoring rubrics to ensure consistency and actively reviewed and corrected complex model outputs in alignment with safety and policy requirements. My efforts were directed at improving evaluator agreement, detecting ambiguous or edge-case scenarios, and optimizing the accuracy of AI systems. • Managed prompt evaluation, response ranking, and preference modeling tasks to enhance LLM outputs. • Created comprehensive annotation standards and procedures. • Conducted detailed safety reviews and ambiguity resolutions for edge cases. • Monitored and analyzed quality signals to drive continuous process improvements.