ML Data Associate at Amazon (LLM training data validation, AI response evaluation, and annotation)
TextSupported LLM training data preparation by processing and validating large volumes of records while improving AI workflows. Evaluated AI-generated responses for quality, factual accuracy, grammar, policy compliance, and instruction adherence to ensure training readiness. Conducted prompt evaluation and performed multimodal annotation tasks while maintaining 99%+ quality. • Processed and validated 5,000+ LLM training records weekly. • Assessed response quality across factuality, policy, grammar, and instruction following. • Performed prompt evaluation and multimodal annotation as part of the labeling pipeline. • Collaborated with Boston and Japan teams on workflow, documentation, and operational efficiency.
2025 - Present