​An event dedicated to exploring all things AI Engineering!

​Event partner: Civo

​​Agenda

​​​​​​🕒 6:00 pm – 6:05 pm
Welcome and Opening Remarks

​🎤 Juraj Stefanic, Technical Content Engineer at Daytona

​​​​​​​​🕒 6:05 pm – 6:20 pm
Talk "Kubernetes Is Not Your Sandbox"​

​🎤 Mislav Ivanda, DevEx Engineer at Daytona

​​​​​​​​Outline:

​Teams are reaching for Kubernetes to run agent sandboxes, and it's the wrong tool. Kubernetes is built to keep things alive and hold them in a steady state. A sandbox is born, forked, and killed before any of that machinery catches up.

The mismatch compounds because the sandbox keeps gaining requirements without shedding any. In eighteen months it went from a fast code-snippet runner, to a stateful box for long-running agents, to ten thousand ephemeral environments that fork for RL rollouts and die in under a second. It has to be all of those at once, a contradiction set no orchestrator was designed to hold.

The cost shows up the moment you measure it. We ran the same 50-action bug-fix trajectory across five stacks and got a 12x spread: 12.9s on the fastest, 161.5s on the slowest. The gap isn't compute, it's lifecycle overhead per action. We name every stack and explain the mechanism behind each number.

​​🕒 6:20 pm – 6:35 pm
Talk "Sovereign AI, End to End"

​🎤 Josh Mesout, Co-founder of relaxAI, CIO at Civo

​​​​​​​​Outline:

​relaxAI was founded in 2023. Since then we’ve been helping regulated organisations and privacy-driven industries take back control. The landscape of AI regulation, risk management, and what is accessible to whom is in a constant state of flux, often rapidly changing overnight. The answer is owning your AI end to end — the compute it runs on, the models you serve, the agents you build, and everywhere your data moves between them. I’ll share what we’ve learned building this with customers, and how you de-risk and save costs across your AI strategy.

​​🕒 6:35 pm – 6:50 pm
Talk "The IDE left the laptop "

​🎤 Dan Bekirov, CEO of Sparkles

​​​​​​​​Outline:

​Opening a pull request is now the easy part. Any agent can sling code into a branch in seconds. The hard part is everything after: proving the change actually works, surviving CI, passing review, and merging without a human babysitting every step.

​We need an IDE for that entire lifecycle on the cloud. Not a text editor in a browser. A control plane that takes a change from first intent through planning, implementation in sandboxes, testing in real environments, review, fix loops, and merge. Testing is not optional glue at the end. Whether an agent runs the suite or a human clicks through a live preview, verification has to live in the same surface as authoring. "Works in the agent box, dies in CI" is a first-class failure mode. So is a clean-looking PR that has never actually run.

​The metric that matters is end-to-end velocity to merge. Not autocomplete. Not "agents write code." Time from intent to a merged, tested, production-ready change.

​​🕒 6:50 pm – 7:00 pm
Talk "Moving towards software factories"

​🎤 Gautham Senthilnathan, CEO and Co-Founder of Calibre

​​​​​​​​Outline:

​Most teams still use coding agents interactively. That works well for building new features, but software spends most of its life in production. This is where software factories come in: systems that continuously turn production signals into tested, reviewable changes.

​We will talk about a real software factory experiment we're running at Calibre called Cepheid.

​Cepheid is a maintenance agent that watches user sessions, combines replays with errors, logs, and code context, then investigates potential issues. When it finds a bug, it can reproduce it, implement and test a fix, and open a draft pull request. We’ll also talk through how we moved Cepheid from our laptops to Daytona sandboxes so it can operate continuously.

​​🕒 7:00 pm – 7:10 pm
Talk "Scaling Production Agents with Adaptive Sandboxing"

​🎤 Alessio Micali, Founder of Polpo

​​​​​​​​Outline:

​We'll give a concise overview of Polpo's adaptive-sandboxing architecture, then demonstrate it. We'll call the same agent twice: a simple request completed without a sandbox, followed by a complex task requiring code and filesystem access that automatically provisions a Daytona sandbox in real time. We'll inspect the trace to show when isolation is activated, how files and results move through the workflow, and how the sandbox lifecycle is managed end to end.

​​​​​​​​​​​🕒 7:10 pm - 8:30 pm

​​​Networking

​With pizzas and beverages

​About event

​This is dynamic gathering for AI enthusiasts, innovators, and professionals to collaborate, share ideas, and explore the latest advancements in artificial intelligence. Whether you're building AI products, researching cutting-edge algorithms, or simply passionate about the field, join us to connect, learn, and drive the future of AI forward.