MLOps -- Understanding Data Drift. Types of Data Drifts and Monitoring…

One of the important functions of MLOps engineers is to monitor the model performance. Data drift causes degradation in the model performance over a period of time. Let's discuss data drift and the steps we can take to detect it in detail. Data drift refers to changes in the data distribution over a period of time. Data drift can lead to poor model performance, because the model is being applied to data that is different from the data it was trained on.