ROC-AUC (Receiver Operating Characteristic Area Under the Curve) is an important metric in machine learning because it measures the quality of a binary classifier's predictions. This is a common topic in data science interviews, so it's helpful to brush up on this concept.