Evaluation and quality: Build golden sets and ground truth, run offline and live evals, choose appropriate metrics (accuracy, precision, recall, F1, task success, hallucination rate), and use results, including LLM-as-judge where validated, to decide what ships and to measure impact. You will own an AI product or major capability within our AI portfolio, define what success looks like, build and develop with the team, and drive it from deeply understanding a business challenge, to prototyping an AI solution, to proving its value, to delivering it through Humana''s enterprise AI governance process.