Hugging Face integrates one-click deep links to Amazon SageMaker Studio for model customization and deployment

Hugging Face and Amazon announced a deep-link integration that allows developers to transition from discovering a model on Hugging Face to a pre-configured Amazon SageMaker Studio environment with a single click. The integration supports two workflows: "Customize on SageMaker AI" and "Deploy on SageMaker AI."

When a developer selects "Customize on SageMaker AI" from a supported Hugging Face model page, they are taken directly to the Model Customization page in SageMaker Studio with the model pre-loaded and ready for fine-tuning. Selecting "Deploy on SageMaker AI" opens the Deployment page in Studio with the model pre-configured for endpoint deployment. The context of the selected model is preserved throughout the transition.

The integration introduces pre-configured permissions for new Studio environments created through this flow. A new managed IAM policy, AmazonSageMakerModelCustomizationCoreAccess, is automatically created and attached, providing permissions for serverless model customization jobs using supervised fine-tuning (SFT), direct preference optimization (DPO), reinforcement learning with verifiable rewards (RLVR), and reinforcement learning from AI feedback (RLAIF). It also supports deployment to SageMaker AI or Amazon Bedrock endpoints.

The integration also surfaces GPU quota visibility directly within the Studio UI during instance type selection for deployment or training. Developers can see which GPU instance types (G5, G6) are available under their account’s current limits without navigating to the Service Quotas page. If a quota increase is needed, they are redirected directly to the relevant Service Quotas page.

The announcement includes a step-by-step walkthrough of the experience. Starting from a Hugging Face model page, developers can select "Customize on SageMaker AI" or "Deploy on SageMaker AI," sign in to AWS (or skip if already signed in), and land directly in the relevant SageMaker Studio workflow with the model pre-selected and the environment pre-configured.