Evaluate Models for Credit-Risk Scoring at Capital One
Scenario
You are building a production-grade credit-risk scoring model (predicting probability of default within a fixed horizon) for Capital One. The model will be used for underwriting decisions and must meet performance, compliance, and interpretability requirements.
Task
Compare logistic regression, random forest, and gradient boosting for credit-risk modeling. For each, discuss pros and cons in this context. Then describe how you would:
-
Evaluate model performance (both discrimination and calibration), including appropriate train/validation splits.
-
Handle class imbalance in defaults.
-
Ensure model interpretability and compliance-readiness.
Include specific metrics (e.g., ROC-AUC, KS), imbalance techniques (e.g., class weighting, SMOTE), and explainability approaches (e.g., SHAP) and how they fit into a regulated credit environment.
Constraints & Assumptions
-
Preserve the scope, facts, inputs, and requested outputs from the prompt above.
-
If the prompt leaves a detail unspecified, state a reasonable assumption before relying on it.
-
Keep the answer interview-ready: concise enough to present, but concrete enough to implement or evaluate.
Clarifying Questions to Ask
-
Clarify the task, data shape, labels, constraints, and evaluation metric.
-
State assumptions behind the math or modeling technique you choose.
-
Connect theory to practical training, debugging, and deployment implications.
What a Strong Answer Covers
-
Correct definitions and formulas where the prompt requires them.
-
A practical explanation of how the method behaves on real data.
-
Trade-offs, failure modes, diagnostics, and mitigation strategies.
-
Evaluation choices that match the product or modeling objective.
Follow-up Questions
-
How would noisy labels, class imbalance, or distribution shift affect the answer?
-
What would you monitor after deployment?
-
Which baseline would you compare against first?
