Amazon SageMaker AI introduces agentic experience for rapid model customization

Amazon SageMaker AI has unveiled a new agentic experience designed to significantly expedite the process of model customization. This development reduces the time required from months to mere days or hours. When constructing an AI solution, customers must delineate their use case objectives and success metrics, prepare data, select appropriate models, and conduct multiple experiments using various models and fine-tuning techniques. Upon identifying an optimal model that satisfies the success criteria, they must then determine the most cost-effective deployment strategy. Throughout this process, customers traditionally manage the substantial task of setting up the necessary infrastructure for training and deploying models.

The newly introduced capability allows developers to leverage natural language interactions with coding agents, thereby streamlining the entire process—from defining use cases to deploying a high-quality model in production. This agentic experience, grounded in SageMaker AI model customization agent skills, offers specialized expertise in fine-tuning tailored to a developer’s specific use case. It facilitates data transformation into required formats, employs comprehensive quality evaluation using LLM-as-a-judge metrics, and provides flexible deployment options to Amazon Bedrock or SageMaker AI endpoints.

Customers can incorporate these skills into any Integrated Development Environment (IDE) such as Visual Studio and Cursor. Developers have the option to collaborate with multiple coding agents, including Kiro, Claude Code, and CoPilot, to optimize popular model families like Amazon Nova, Llama, Qwen, and GPT-OSS. This experience produces reusable, editable code artifacts, enhancing transparency, reproducibility, and automation through integration into AIOps pipelines.

To install SageMaker AI skills, developers can use the sagemaker-ai agent plugin in their preferred IDE. These model customization skills are also pre-installed in SageMaker Studio Notebooks, accompanied by the Kiro coding agent. Users can subscribe to Kiro, open the chat window in Studio Notebooks, and begin interacting with the agent to develop their workflow. The platform supports advanced customization techniques such as supervised fine-tuning for instruction tuning, direct Preference Optimization for tone and preference adjustments, and Reinforcement Learning for scenarios requiring verifiable correctness.

For further information on model customization using the AI agent experience in Amazon SageMaker AI, users are encouraged to consult the SageMaker model customization documentation.