Context

LendWise uses a binary classification model to predict whether a personal loan applicant will default within 90 days. The model was trained on 2023 application data and deployed in January 2024, but the risk team now suspects significant data drift after a new customer acquisition channel launched in Q3.

Current Performance

The Problem

The model is underestimating risk in the most recent population while business conditions and applicant mix have changed. The head of credit wants to know how you would validate whether the model is still reliable, determine whether the issue is covariate drift, concept drift, or calibration failure, and recommend next steps.

Requirements

1. Explain how you would validate the model after detecting drift in the input data.
2. Identify which metrics suggest the model has degraded and why.
3. Propose a structured validation plan to separate data drift from target/concept drift.
4. Recommend actions for short-term risk mitigation and longer-term model improvement.

Constraints

• Loan approvals cannot pause during validation.
• False negatives are costly because missed defaults increase charge-offs.
• Retraining requires at least 3 weeks due to label delay and compliance review.