• Own and manage the backlog for ML-driven biometric and document verification capabilities. • Translate fraud, identity, and customer requirements into clear and actionable ML work items. • Partner closely with ML engineers and data scientists to refine problem statements into feasible deliverables. • Define acceptance criteria that reflect real world performance, not just offline model metrics. • Participate actively in model design discussions, prioritization, and tradeoff analysis. • Support model lifecycle activities including training, evaluation, deployment, and retraining. • Ensure monitoring, drift detection, and feedback loops are incorporated into delivery plans. • Partner with agent operations and data teams on labeling strategy and data quality. • Incorporate fraud patterns and adversarial thinking into backlog prioritization. • Work closely with engineering, fraud, compliance, legal, and customer teams.