For decades, code separated human intent from execution. Today, software can execute directly from intent. Just as cloud engineering used Infrastructure as Code (IaC) to shift from imperative scripts to declarative state, modern AI extends this evolution through Infrastructure as Intent (IaI). This changes how we build the AI Data Stack. This session explores what happens when human expertise combines with structured intent to mentor AI systems. Grounded in my research on Intent-Driven Development (IDD) and lessons from building agentic tools, we will examine how to structure intent, measure system leverage using the Intent Compression Ratio (ICR), and turn transient prompts into durable Skills. We spent decades optimizing stacks for syntax and code. The next challenge is engineering intent.