Senior Data Annotation Lead at Apex Intelligence Inc.
Led RLHF preference ranking annotation workflows to support large language model fine-tuning serving millions of users. Implemented active learning loops to prioritize high-uncertainty samples and optimize the annotation queue per training cycle. Established inter-annotator agreement measurement using Cohen's Kappa to improve label consistency and reduce rework. • Coordinated annotation teams (22 annotators, 3 QA analysts) across multiple LLM and CV initiatives • Improved helpfulness and safety benchmarks through RLHF-driven data work • Reduced turnaround time and rework via IAA scoring system • Refined labeling schemas for low-performing edge classes to boost precision