Data Annotation Generalist
I assess model-generated responses using detailed rubrics focused on accuracy, instruction-following, reasoning quality, clarity, and safety. I compare outputs across a wide range of prompts, identify strengths and weaknesses in each response, and provide concise but well-supported justifications to ensure evaluations remain consistent, objective, and aligned with project guidelines.