Core

Data and Concept Drift

A model can lose accuracy two ways: the inputs it sees start looking different (data drift), or the true mapping from inputs to outputs changes underneath it (concept drift). The fix differs, so the FDE skill is diagnosing which one you have before reaching for a retrain.

a free account unlocks the core curriculum tier · no card

RELATED CONCEPTS

Model Monitoring Observability for AI Systems Golden Datasets and Eval Sets CI/CD for Models

LESSONS THAT TEACH THIS

The FDE EngagementEvery dashboard is green and the system has been getting worse for a month25 min→The FDE EngagementA drift alarm is a reason to investigate before you retrain28 min→

PRACTICE THIS IN REAL QUESTIONS

MLOps & ML EngineeringHow is CI/CD for ML models different from traditional DevOps CI/CD?→ML System Design (Product)Build a fraud-detection model for payments.→System Design & Production EngineeringDesign observability for an LLM application. What do you log, trace, and alert on, and how is it different from normal services?→MLOps & ML EngineeringYour pipeline retrains automatically. How do you decide whether the new model replaces the old one?→MLOps & ML EngineeringDesign an automated retraining pipeline. What should trigger retraining?→MLOps & ML EngineeringWhat is point-in-time correctness, and how do you avoid leakage in continuous retraining?→