Experience defining and implementing presentation attack detection and an adversarial threat model with a fraud-first mindset across modalities, including voice replay/deepfake resistance and call audio injection risks (noise and device/line variability), face liveness and deepfake/face-swap and injection defenses (with device integrity signals), and iris/palm considerations (sensor trust, presentation attacks, and deployment constraints such as specialized hardware and environmental requirements). Experience engineering end-to-end identity lifecycle for biometrics-secure enrollment (proofing strength, step-up, dedupe, re-enrollment rules), strong binding to the correct account/device/session with change handling (new device, biometric drift/injury/aging), secure recovery that avoids lockouts, and clear consent/transparency with data-minimization practices.