AI Data Annotation Specialist (Part-time), Talent AI
Image
RLHFProvided multilingual data annotation and conducted quality review for AI training datasets at Talent AI, strictly following annotation guidelines. I specialized in LMH image-text cover (图文封) annotation, where I selected suitable images and crafted challenging questions based solely on visual information in the image. The goal was to create difficult questions designed to test the AI model’s limits, while providing detailed reasoning guidance and error corrections. These tasks focused on improving the model’s understanding of complex topics such as humanistic history, architectural spaces, and route planning. I also performed long text analysis and image-text matching projects. My work emphasized ensuring label consistency, accuracy, and high data quality to effectively support AI model training and evaluation. I adapted quickly to evolving annotation standards and consistently met tight deadlines. Key Contributions: Designed challenging questions for LMH image-text cover tasks by selecting images and creating questions that rely exclusively on visual content, aiming to rigorously test and improve the AI model’s comprehension of humanistic history, architectural space, and route planning. Provided detailed reasoning guidance and error correction to strengthen model performance. Annotated and analyzed long text content for AI training datasets. Performed quality checks and validation to maintain high annotation accuracy. Collaborated with project teams to improve labeling efficiency and overall data quality.