With tools like Claude Code and Copilot now writing significant chunks of our Ember components and templates, how do we make sure we aren't just shipping bugs faster?

In this talk, I’ll share how we can keep our Ember apps stable when we’re leaning on AI to help us code. We’ll look at practical ways to wrap our agent-assisted workflows in solid "guardrails" — specifically focusing on how to level up our automated testing. We’ll walk through:

• Coverage that counts: Moving beyond raw lines of code to testing actual component behaviour.
• Visual Regression: Automating the "eye test" to catch UI shifts before they reach the browser.
• Custom Linting: Using custom ESLint rules to bake our team’s Ember conventions directly into the AI’s workflow.

If you’re worried about AI making a mess of your codebase, come by to swap stories and pick up some patterns for keeping our Ember apps as reliable as ever.

David McNamara

In this talk, I'll tell you more about my experience as an Ember developer who wanted to create a browser extension.

We'll take a look at how I used broccoli 🥦 to build my extension for Chrome and how I switched to vite with plugins to target all the browsers...

KB

Karel Bousson

Dependency resolution in node projects is a complex beast at the best of times. Peer dependencies, competing module standards, bundlers and transpilers can seriously muddy the waters. Things can get especially hairy during addon development, where you'd like your app to resolve your local version.

In this talk we'll try to build an intuition for what is going on underneath, so that we can understand why things go wrong.

Arne Bertrand

After a brief stint in corporate finance, I was lucky to discover my true passion and able to make a full career switch to software development. I especially love building complex GUI software, as it provides a challenge right in the middle of UX design, performance-critical algorithms, and software architecture.

LLMs, agents and AI changed the way we develop code and products. It seems that the future is now. How fast-paced teams work these days and what the future may hold.

I'll prove that saying that Ember is not ready for LLMs is a false claim.

Tomek Nieżurawski