AI-assisted data development with Kiro and SageMaker Unified Studio

This article demonstrates how to connect AI coding assistants like Kiro, VS Code, and Cursor to Amazon SageMaker Unified Studio for governed, AI-assisted data development.

• Connect local IDEs to SageMaker Spaces via secure SSH tunnel using AWS Toolkit v4.1.0+
• SageMaker automatically generates AGENTS.md and smus-context.md steering files providing AI context about project environment and data catalog
• Configure MCP servers (smus_local and aws-dataprocessing) to give AI agents direct access to AWS Glue Data Catalog and Athena queries
• Use natural language prompts to explore data, run analytics, and generate code while maintaining data governance and access controls
• Best practice: explore data first before building to reduce LLM hallucinations and improve code quality on first pass
• Supports Jupyter notebooks with language and connection selectors matching SageMaker JupyterLab experience

This integration enables productive agentic AI development without compromising organizational data governance, security controls, or compliance requirements.