AWS adds SageMaker AI inference-optimization skill for coding agents

AWS introduced the aws-ai-ml skill for optimized generative AI inference on Amazon SageMaker AI through the Agent Toolkit for AWS. The company says it gives coding agents guidance for inference optimization and benchmarking. Examples named by AWS include Kiro, Claude Code, and Codex; the skill can also plug into agents that support the Model Context Protocol.

Users can describe what they want in natural language, and the agent may ask clarifying questions before generating code. The skill can benchmark endpoints, recommend deployment configurations, compare performance runs, and produce executable SageMaker Python SDK v3 code. AWS says the generated code is grounded in real benchmarks and measured performance data, and can be reviewed, modified, and run in a user's own environment.

The skill can be installed locally through the Agent Toolkit for AWS or used in an Amazon SageMaker Studio JupyterLab space. Its benchmarking workflow fits into SageMaker AI's broader inference-recommendation work: AWS documentation describes evaluating configurations on real GPU infrastructure and returning performance metrics.