CUA Tutor · xAI
Produced and curated high-quality labeled datasets for Grok's Computer Use Agent (CUA) and web-based workflow training and evaluation. Designed, annotated, and reviewed complex multi-step web-based interactions involving browser navigation and form handling. Evaluated agent and model outputs for accuracy, reasoning, consistency, and alignment with objectives. • Curated data spanning real-world web environments and automation contexts. • Performed human-in-the-loop feedback processes to improve AI agent performance. • Applied both technical annotation and evaluative review steps. • Supported systematic enhancement of model reliability and task completion.