Run interactive IDEs on Amazon EKS with SageMaker AI to power up your AI workflows

The Amazon SageMaker AI Spaces add-on for Amazon EKS enables data scientists to run interactive IDEs like JupyterLab and Code Editor on the same cluster as their pipelines, eliminating the need to switch to a standalone JupyterHub deployment or local laptop. This solution can increase GPU utilization by up to 30 percent and reduce costs by avoiding the need for an always-on GPU environment. The add-on can be set up in about 5 minutes, compared to the 3-5 days it typically takes to stand up a standalone JupyterHub environment. The solution runs on a single EKS cluster in three layers: network and access, cluster routing, and compute and storage. For engineers building AI systems, this means they can streamline their workflow and improve productivity by having all their tools and resources in one place.

⚡ Key Takeaways

• The Amazon SageMaker AI Spaces add-on for Amazon EKS can be set up in about 5 minutes.
• The solution runs on a single EKS cluster in three layers: network and access, cluster routing, and compute and storage.
• The add-on uses AWS Load Balancer Controller, Traefik, and Auth middleware to manage access and routing.
• The solution uses Amazon Elastic Block Store (Amazon EBS) CSI driver and Amazon Elastic File System (Amazon EFS) or Amazon FSx for persistent volumes and shared storage.
• The add-on must be version 0.1.4 or later to support Amazon SageMaker AI Spaces.

The Amazon SageMaker AI Spaces add-on for Amazon EKS has a significant impact on engineers shipping production AI today, as it streamlines their workflow and improves productivity by providing a managed JupyterLab and Code Editor environment on the same cluster as their pipelines. This can lead to increased GPU utilization and reduced costs.

✅ Practical Steps

1. Install the SageMaker AI Spaces add-on on an Amazon EKS cluster using Helm v3.
2. Set up the supporting add-ons and IAM roles, including the AWS Load Balancer Controller and AWS Key Management System (AWS KMS) encryption key.
3. Create a Space and access it through a presigned URL in the browser and from VS Code over SSH-over-SSM.
4. Review how to move your team to OpenID Connect (OIDC) sign-in with Amazon Cognito.

Want the full story? Read the original article.