AI Trainer
As an AI Trainer at OutlierAI, I evaluated, wrote, and refined AI prompts and responses to enhance AI system accuracy. My work focused on rating AI-generated responses for accuracy, safety, and quality in order to train and improve underlying large language models. I identified nuances and errors in AI outputs, contributing to the evolution of conversational AI performance. • Evaluated and scored AI-generated textual responses for correctness and safety. • Wrote and edited prompts and target completions for supervised fine-tuning tasks. • Provided detailed feedback and recommendations on AI behaviors and errors. • Collaborated with team members to identify edge cases and improve model reliability.