Skip to main content

Platform Integration

Since AI-Link provides programmatic interaction with the Semantic Layer via Python, data professionals can use it with any Python-based libraries, feature stores, AutoML products, and ML frameworks that they choose.

AI-Link also supports any warehouses, catalogs, or data lakes that AtScale itself supports. Please refer to the AtScale documentation for a full list of supported platforms.

As of Version 2.6.0, AI-Link supports OAuth and SAML authentication.

Yes – AI-Link and Snowpark both allow Pythonic interaction with data at the warehouse level, so use of one can very well complement the other. For one example, the business-vetted features that AI-Link serves can be used to train a model locally. A user can then use Snowpark to create a user-defined function (UDF) in Snowflake that renders predictions via the locally-trained model.

Yes – AI-Link works with providers like DataRobot, Dataiku, Sagemaker, and H2O to facilitate your ML workflows. It provides a single source of business-vetted features, eliminating arduous data pipeline management and automating feature creation and engineering.

Yes – DataFrame output from AI-Link can be ingested by ML pipeline platforms such as Kubeflow and Airflow.