Amazon Web Services (AWS) launched its SageMaker Operators for Kubernetes that uses the Kubernetes Operators model to more tightly couple the SageMaker machine learning (ML) platform with their Kubernetes workflows.

The SageMaker Operators for Kubernetes product allows users to tap into data housed within SageMaker to populate a Kubernetes-controlled container or cluster of containers. This can be done at the scale needed to support ML-based services and does not require the data scientist or developer to re-write code.

The integration uses the Kubernetes Operators model that allows a user to natively invoke custom resources and automate associated workflows of pre-configured application-specific or domain-specific logic and components. SageMaker can be used as one of these custom resources.

The Operators model was originated by CoreOS as a controller that runs Kubernetes for a particular application. It does this by using the Kubernetes API to handle the creation and management of application instances. The concept is targeted at distributed applications and allows for the scaling of instances as needed.

Initial SageMaker Operators include Train, which helps to cut training costs; Tune, which automates hyperparameter optimization; and Inference, which can handle the autoscaling of container clusters that are spread across multiple availability zones. Those initial zones include AWS’ US East (Ohio), US East (N. Virginia), US West (Oregon), and EU (Ireland).

AWS unveiled SageMaker at its annual re:Invent show in 2017. It’s a fully managed, end-to-end ML service that helps developers build ML models at scale.

The cloud giant at that event also launched its Elastic Kubernetes Service (EKS), which (finally) offered a managed Kubernetes service running on top of AWS. This allows users to deploy and manage containerized applications while staying within the warm embrace of AWS.

AWS this week also linked its SageMaker product into a newly developed Adlink AI at the Edge platform that is targeted at industrial use cases that can benefit from artificial intelligence (AI) at the edge. It combines AWS’ SageMaker and Greengrass edge platform with Intel’s OpenVINO toolkit, which includes accelerators and streamlines deep learning workloads across Intel architecture, and edge company Adlink’s Edge software suite.