AI Trainer and Evaluator Generalist Expert
As an AI Trainer and Evaluator Generalist Expert at Mercor, I write realistic prompts, annotate and categorize data, and evaluate AI model responses for accuracy, reasoning, and user intent. My duties include rubric-based grading of AI outputs, application of predefined taxonomies, and consistently flagging dataset inconsistencies. I contribute to the overall reliability, safety, and performance improvement of AI systems through structured annotation work. • Write prompts and responses covering general topics for text-based AI training • Evaluate and compare AI-generated outputs using rubrics and guidelines • Annotate large volumes of text and other data types according to detailed instructions • Identify ambiguities and errors, ensuring quality and consistency across datasets.