Skip to article frontmatterSkip to article content
Site not loading correctly?

This may be due to an incorrect BASE_URL configuration. See the MyST Documentation for reference.

Monitoring & Drift Detection

MLOps doesn’t end with deployment. You must monitor the model’s performance in the real world.

1. Logging Production Metrics

You can use MLflow to log metrics from your production inference server to track accuracy over time.

2. Drift Detection

If the distribution of your input data changes (Data Drift), your model’s accuracy will likely drop. Use MLflow to log histograms of input features and compare them with the training distribution.