End users of Snowflake can now seamlessly integrate their data into Azure Machine Learning (Azure ML) for training this data for AI models. Azure ML’s functionality has also been extended.
The integration with the large data warehouse platform by Microsoft allows users, such as data scientists, to import Snowflake data directly, with a single command, into Azure ML for further processing. Without having to rely on third-party solutions or by engaging real data specialists to do so.
The integration brings data scientists a number of benefits. These include improved collaboration between data scientists, more time savings, a simplified workflow, more flexibility and more traceability of the data, especially in the case of re-training this data for audits of the developed AI or ML models.
By the way, Microsoft indicates that this integration capability also applies to cloud-based storage services such as AWS S3.
Other new functionality Azure ML
The tech giant further announced several improvements to Azure ML functionality. In public preview, lifecycle management capabilities are now available for managing imported datasets in an Azure ML datastore or a so-called ‘hosted on behalf of ‘ (HOBO) environment. Mainly for data imported via the CLI and the SDK.
In addition, new tracking tools were also presented. These should help companies manage AI or ML training tasks. A customizable list of training tasks has also been introduced. Last, but not least, functionality has been added that merges measurement data and images of training projects, lets you take notes and create and save custom views.