Most frontend developers use AI agents like a fancy autocomplete... and then wonder why the output is mediocre. The difference between "meh" and "mind-blowing" isn't the model; it's how you engineer the context around it.
In this talk, I'll show you how to treat AI agents like a junior developer joining your team. You'll learn how to onboard them with rules and guides, manage their working memory through context engineering, and extend their capabilities with MCP tools that let them browse, test, and even verify accessibility (WCAG) on their own.
You'll walk away with a repeatable framework you can apply tonight to make AI agents actually useful in your frontend workflow; not just impressive in a demo.

Target Audience: Frontend developers (all frameworks) who are using AI coding agents but not getting great results. No prior AI/ML knowledge required.
Key Takeaways:
How to write rules and guides (markdown files) that shape agent behavior before a single prompt is written.
What context engineering is and why it matters more than prompt engineering for real-world frontend work.
How MCP tools turn AI agents from text generators into agents that can browse, test, and verify accessibility (WCAG).
A repeatable framework for "onboarding" AI agents into any frontend codebase.
Preferred session duration: 20-30 minutes.