dltHub has launched a free, self-paced 1-hour course called 'Agentic Data Engineering with dltHub' covering how to build production-grade data pipelines using AI agents. The course spans five lessons across four phases: defining outcomes, building ingestion pipelines, modeling data, and shipping/monitoring. It introduces the AI Workbench — an instruction set that guides agents (Claude, Cursor, Copilot) through a 12-step workflow with guardrails around credential handling, schema validation, incremental loading, and deployment. The course targets data and analytics engineers who already use AI coding tools daily and want pipelines that outlast the initial prompt, not beginners. Key differentiators include metadata flowing between pipeline steps so agents maintain context, and a propose-verify-enforce pattern that prevents agents from skipping critical checks.
Table of contents
What "it runs" doesn't cover Link iconThe workflow runs on metadata Link iconThe workflow, end to end Link iconWhere the lessons fit Link iconWho it's for Link iconWhat does it do for you? Link iconHow to start Link iconSort: