Amazon SageMaker AI introduces agent-guided model customization

AI Impact Summary

Amazon SageMaker AI now offers an agentic workflow for model customization, streamlining the traditionally complex process of fine-tuning models. This new experience leverages a coding agent, like Kiro or Claude Code, guided by pre-built ‘Skills’ that automate tasks from data preparation and technique selection to evaluation and deployment. This shift reduces the technical burden and accelerates experimentation by providing specialized AWS and data science expertise directly within the JupyterLab environment.