This meetup is for developers and platform teams who want to understand what it actually takes to move from AI experimentation to production-ready workflows. We’ll focus on how teams are redesigning their development workflows to work with AI, not just adding tools on top. That includes how to structure workflows, manage agents, and introduce the right level of control without killing developer velocity. No hype. No tool comparisons. Just a clear look at why most AI workflows fail and what actually works at scale.

Who this is for:
- Beginner → intermediate software developers
- Platform engineers and DevOps folks supporting dev teams
- Anyone using (or wanting to use) AI tools beyond simple autocomplete

Agenda:
* 5:30pm~6:00pm: Checkin, Food and Networking
* 6:00pm~6:15pm: Welcome/community update
* 6:15pm~8:00pm: Tech talks and Q&A
* 8:00pm~8:30pm: Open discussion & Mixer

Tech Talk: Building Functional Dev Environments for Humans and Agents
Speaker: Ben Potter, Head of DevRel. @ Coder
Abstract: ​Your laptop breaks, the build gets complicated, or a new engineer joins the team, and suddenly the development environment becomes part of the problem. Maybe you have your setup perfectly dialed in, but nobody else can reproduce it. Now add AI agents that need the same environment to actually build, test, and ship code. The gaps get bigger fast. This session looks at practical patterns for managing inner-loop development environments across everything from solo projects to enterprise-scale multi-repos and monorepos. We’ll get into what makes an environment reproducible, how to keep humans and agents working from the same foundation, and why the developer experience becomes increasingly important as software development changes in 2026.

Tech Talk: Don’t Blow Your Token Budget Being an Agent Psychologist
Speaker: Roni Dover, Head of AI Labs, JetBrains.
Abstract: ​A file delete that should have cost nothing came to $30, all from context that nobody needed. The bill is where you find out, but it is not where the problem started. Agents go wrong in ways you can learn to recognize. It can be an early wrong assumption that never gets revisited, so every step after it is paid for twice. Work reported as finished when nothing was verified. A run that stops halfway and asks permission instead of doing the thing. A real defect noticed briefly and then filed away to be forgotten. None of this looks like failure while it is happening, which is why the number at the end is a surprise. This session walks through those patterns with real trajectories from live coding sessions and the costs they produce, then gets into what can catch mid-run, and how. The agent almost always tells you, in its own words. You just have to know how to listen.

Venue:
Coder, 520 E Oltorf St, Austin, TX 78704