AWS has detailed how machine learning teams can run interactive development environments directly on Amazon EKS using the SageMaker AI Spaces add-on. The setup provides managed JupyterLab and code editor spaces on Kubernetes clusters that teams already operate.
The workflow is aimed at organizations that want browser-based development while keeping workloads close to their container infrastructure. AWS describes browser access, VS Code connections over SSH-over-SSM, and a move toward OpenID Connect sign-in through Amazon Cognito.
For ML platform teams, the appeal is operational consistency. Developers can get familiar notebook and coding tools without creating a separate environment outside the cluster. The limitation is complexity: this is most relevant to teams already comfortable with EKS, identity configuration, and cloud networking, not to someone looking for a simple hosted notebook with minimal setup.