Freelance AI Trainer & Data Annotator
I completed fine-grained video action annotation tasks for AI training datasets, labeling human behaviors using structured verb-object frameworks. My duties included evaluating LLM-generated content and providing preference labels for RLHF pipelines. I consistently maintained high annotation accuracy and participated in quality assurance and red-teaming exercises. • Labeled video data to capture human actions for model training • Evaluated LLM outputs on accuracy, helpfulness, and safety • Maintained 98%+ agreement in quality-focused annotation tasks • Contributed to creative-domain red-teaming for model robustness