Interview · flashcards

MLOps & Data Quality

Data drift, class imbalance, monitoring, CI/CD for ML and everything that keeps models alive in production.

easymonitoringdata-quality

What is data drift and how do you detect it?

Click, press Space or use the button below to reveal the answer

Interview-style answer

$Data drift is a change in the input feature distribution P(X) over time compared to training data.$
$Detect it with statistical tests per feature (KS test for numeric, chi-square for categorical), Population Stability Index (PSI > 0.2 = significant drift), or distance measures (JS divergence, Wasserstein).$
$Monitor drift in production dashboards and alert when thresholds are crossed.$
$Drift does not always hurt performance — measure impact on output and label if available.$

Click again to flip back

1 / 215

Shortcuts: Space/Enter flip · ← / → navigate · K = known · R = review