AI Language Trainer, Curacel
TextEvaluate and rank Large Language Model (LLM) responses to ensure alignment with human values for helpfulness, accuracy, and safety while achieving a 98% quality assurance score. Create high-quality, linguistically complex prompts paired with gold-standard responses to train models for advanced English reasoning, summarization, and creative writing. Provide detailed linguistic feedback to reduce hallucinations and improve syntactical coherence, including identifying edge cases with machine learning engineers. • Response ranking and quality evaluation for LLM outputs • Prompt + response writing for SFT-style training data • Linguistic error and coherence feedback for model improvement • Safety-focused red teaming targeting nuanced cultural contexts