AI Operations & Model Safety Specialist (RLHF Data Labeling & Model Evaluation)
I developed high-quality training datasets and audited 'Chain-of-Thought' reasoning to enhance model accuracy and safety for RLHF projects. I thoroughly reviewed high-risk and ambiguous AI outputs, making defensible decisions on safety, misinformation, and intellectual property issues. I collaborated and contributed to the refinement of review guidelines, process improvements, and delivered structured reports to support risk governance and data quality efforts. • Conducted escalation case investigations to improve model consistency and reduce repeated issues. • Flagged emerging risk patterns and communicated them to engineering and policy teams. • Fine-tuned Large Language Models (LLMs) using human feedback loops and rating workflows. • Managed multiple priority workflows and coordinated with cross-functional teams to resolve labeling and training process issues.