AI Trainer / AI Response Evaluator - Data Annotation & Handshake AI (AI response quality evaluation)
TextEvaluated AI responses for accuracy, instruction following, tone, clarity, and overall task completion quality. Created prompts, test cases, and edge-case scenarios to expose model weaknesses and failure modes. Reviewed model outputs that included JSON formatting and simple logic such as functions/methods and support-style problem solving, then provided written improvement guidance. • Assessed response quality across multiple rubric dimensions (accuracy, instruction following, tone, clarity, completion). • Designed evaluation prompts and edge cases to measure robustness. • Performed structured review of JSON/function-calling-style outputs and reasoning responses. • Supplied written feedback explaining errors, preferred outputs, and areas for improvement.