Senior AI Trainer & Data Annotation Lead (DataMind AI Labs)
Designed and implemented RLHF pipelines for two production-grade LLMs, leveraging human feedback to improve response quality. Led large-scale NLP and computer vision dataset annotation workflows with standardized QA to maintain high inter-annotator consistency. Authored guidelines and rubrics to reduce labeling errors and accelerate onboarding for new annotators. • RLHF feedback generation and dataset refinement • QA rubric development and error-rate reduction • Inter-annotator agreement monitoring using Cohen's Kappa • Instruction-tuning edge-case prompt/response pair creation