Generalist Expert - Mercor
Supported AI model training by synthesizing and organizing large volumes of multimodal datasets using provided taxonomies and annotation guidelines while maintaining high accuracy and consistency. Applied rubric-based frameworks to produce structured outputs and improve reliability of downstream AI systems. Actively identified ambiguities and edge cases, providing feedback to strengthen data quality and guideline clarity for future iterations. • Applied predefined taxonomies and annotation frameworks to text, image, and multimodal data • Maintained 95%+ annotation accuracy and consistent output quality • Flagged inconsistencies, ambiguities, and edge cases to improve dataset reliability • Worked independently in a remote setup while meeting tight deadlines and reporting progress