AI Training and Annotation Tasks in Medical Contexts
As a Freelance AI Trainer on the Outlier AI Platform, I engaged in advanced workflows for model evaluation and improvement in the clinical and STEM domains. My responsibilities included prompt engineering, response evaluation, and rubric-based scoring, focusing on accuracy and quality. I analyzed model failures and provided structured corrections to enhance AI reasoning and safety. • Evaluated AI model outputs for reasoning integrity in clinical contexts. • Applied RLHF and RLVF methods for performance optimization. • Designed prompts to test diagnostic and multi-step reasoning skills. • Produced corrections and feedback aligned with real-world clinical decision processes.