Hands-on Workshop: Building AI pipelines with Apache Airflow®

AI-powered workflows are becoming a core part of modern data platforms, yet teams often struggle to move LLMs, embeddings, and agents from prototype to production. Airflow and the common AI provider address these challenges with purpose-built functionality for LLM tasks, toolsets, and multi-step agents, plus orchestration features like human-in-the-loop approvals and asset-aware scheduling.

In this 90-minute, virtual workshop, you'll build an end-to-end AI pipeline that analyzes customer reviews, routes them to the right team, finds similar complaints, and drafts personalized responses for human review, all in your browser with no local setup required.

You will learn how to:

• Build AI tasks for structured data extraction and branching using @task.llm and @task.llm_branch
• Orchestrate multi-step agents with tools using @task.agent
• Use advanced Airflow 3 capabilities such as human-in-the-loop approvals and asset-aware scheduling

All examples use a realistic scenario at AstroTrips, a fictional interplanetary travel company, so the patterns are easy to follow and ready to adapt to your own workflows.