• Own model evaluation and quality standards • Take full ownership of all evaluation systems, validation sets and quality metrics • Establish and maintain “golden sets” for every product as the source of truth for quality and accuracy • Drive measurable accuracy improvements • Lead accuracy initiatives across all products, improving precision, recall, and overall model performance • Define and track AI performance metrics and client SLAs as we scale • Automate quality workflows • Identify manual QA and data entry processes and redesign them through automation • Reduce human intervention by increasing model confidence, reliability, and repeatability • Build scalable evaluation pipelines • Create repeatable processes to evaluate pilots and active customers quickly • Ensure accuracy metrics remain high as customer volume and product surface area grow • Close the real-world feedback loop • Capture and analyze real-world model errors • Prioritize accuracy issues and partner with Engineering to drive continuous model improvement