AI Trainer / Content Evaluator
As an AI Trainer and Content Evaluator at Scale AI, I evaluated and ranked LLM-generated responses for tone, clarity, fluency, factual accuracy, and audience appropriateness. I provided detailed preference feedback and justifications that contributed to fine-tuning AI models through RLHF pipelines. My role involved crafting diverse prompts and identifying nuanced language issues in AI outputs. • Achieved over 97% inter-annotator agreement scores. • Collaborated with QA leads to clarify guidelines and update editorial style guides. • Specialized in creative writing, conversational, persuasive, and general knowledge AI scenarios. • Worked independently in a fully remote and asynchronous environment.