Supervisory Data Scientist; AI​/ML

Duties

Help

The division's independent scientific evaluation authority for all enterprise AI/ML capabilities by establishing standardized Test, Evaluation, Verification, and Validation protocols.

Directs the generation of Independent Validation and Verification reports that document critical statistical metrics such as ROC-AUC, F1-scores, and precision/recall distributions.

Oversees the branch's continuous monitoring protocols for models already in production, including defining mathematical drift thresholds and automated alerts, ensuring deployed models do not degrade or develop bias over time.

Collaborates with the J6

DAA branch to ensure automated validation scripts are successfully integrated directly into the enterprise machine learning operations (MLOps) continuous Integration and continuous deployment pipelines.

Directs the branch's adversarial testing frameworks, such as data poisoning, model inversion, and evasion attacks, to proactively identify and mitigate vulnerabilities in AI systems.