Senior AI Content Specialist | DataFlow Solutions
Performed RLHF workflows to rank and improve model outputs based on helpfulness, honesty, and harmlessness. Led evaluation and iteration cycles to increase response quality for downstream user experiences. • Defined criteria and preference signals aligned with RLHF reward goals. • Used prompt engineering to probe model behavior for coding, logic, and creative reasoning. • Coordinated large-scale labeling and feedback collection through a team of annotators. • Partnered on bias mitigation using targeted adversarial red-teaming.