AutoML

AutoML is a Databricks ai / ml capability, introduced May 2021.

Automated model selection that applies machine learning to a dataset by searching for the best algorithm and hyperparameter configuration on your behalf.

• Its whole pitch was skipping the boring part of machine learning, generating models from a dataset to shorten the path to production. 🔗
• Its GA was announced against a runtime version rather than a date: generally available in Databricks Runtime 10.4 LTS ML and above. 🔗
• It has quietly been evicted from the runtime it was pinned to - from Databricks Runtime 18.0 ML onward it is no longer a built-in library and has to be installed like any third-party package. 🔗

⚠ Limitations: In Databricks Runtime 18.0 ML and above AutoML is not included as a built-in library, so it must be installed from PyPI; no libraries beyond those preinstalled in Databricks Runtime ML should be added to the cluster, network ports 1017 and 1021 must be open, and standard (formerly shared) compute access mode is unsupported. 🔗

Category

AI / ML

Introduced

May 2021🔗

Also known as

Databricks AutoML, databricks-automl-runtime

Verified

2026-08-20