Three features now available in GA enable you to streamline your network isolation experience, deploy your model locally prior to production, and automate the ML lifecycle without accruing extra compute hours or requiring manual intervention.
Managed Network Isolation: You can now execute a faster, streamlined workspace setup, and operate with automated network configurations and simpler architecture.
Inference Server : You can now debug scoring scripts locally to diagnose and resolve issues prior to production deployment.
View statuses of AzureML jobs directly in Azure Pipelines : You can now avoid unnecessary compute cost accrual by eliminating the need for continuous polling.