You've shipped a model that looked strong in validation, but it is underperforming after deployment. You need to figure out why the offline results did not hold up in the real world.
Tell me about a time your model performed well in validation but failed in deployment. How did you debug it?