E-commerce Product Classification - 50K Images
Evaluate and rank AI-generated responses for large language model training. Review 200+ conversation pairs daily across diverse topics including technical support, creative writing, and factual information retrieval. Rate responses on accuracy, helpfulness, harmlessness, and coherence using detailed rubrics. Provide written feedback on model outputs to improve alignment. Contribute to prompt engineering by crafting high-quality instruction-response pairs. Maintain 95%+ inter-rater agreement scores with quality assurance team.