Your question is Explainability and Compliance. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
What steps do you take to ensure your AI models are explainable and compliant with institutional standards?
Explain how you would select an interpretable model or explain a complex one, validate global and individual predictions, assess fairness, document assumptions and limitations, and support independent model validation. Address the full lifecycle from data preparation and training through deployment, monitoring, change control, and retirement. Use Citi's Model Risk Management expectations where relevant, but distinguish mandatory policy requirements from technical best practices.