​An event dedicated to exploring all things AI Engineering!

​Event partner: Descope

​Agenda

​​​​​🕒 5:30 pm – 5:40 pm
Welcome and Opening Remarks

​🎤 ​​​​Asako Hayase, Pacer, Community Ambassador at Daytona

​​​​​​🕒 5:40 pm - 6:00 pm
Keynote "Auth Best Practices for AI Agents and MCP Servers"

​​🎤 Kevin Gao, Head of DevRel at Descope

​​​Outline:

​Every organization has an agentic AI / MCP project, but how many of these are in production? And if they aren't in production, how likely is it that authentication, access control, and agentic identity concerns are the reason? In this session, Descope's Head of DevRel Kevin Gao will share real-world MCP use cases, common agentic identity challenges, and practical identity tips to secure AI agents and MCP servers.

​🕒 6:00 pm - 6:20 pm
Keynote "Running AI coding agents in cloud sandboxes"​

​​​​🎤 James Murdza, Pacer, Community Ambassador at Daytona

​​​​​​Outline:

​​​​Claude Code, Codex, OpenCode are three of the top CLI-based coding agents today. We can make coding agents more powerful with cloud sandboxes—adding safety, scalability and repeatability. James will give demos of these agents running in the cloud and walk through the technical details.

​​🕒 6:20 pm - 6:30 pm
Lightning Talk "Claude wants some guardrails"

​​🎤 James Cowling, Co-founder and CTO of Convex

​​​Outline:

​Programming is dying. Engineering is in more demand than ever before. Gemini isn't going to build Google no matter how many tokens you give it - the limited resource in the software industry is design and architecture. We'll talk about what kind of architectural guardrails can be used so that LLMs produce not just good code but good systems, that can scale, can be maintained, and can be understood by engineering teams.

​​🕒 6:30 pm - 6:40 pm
Lightning Talk "Space Machine Sandboxes"

​​🎤 Kwindla Hultman Kramer, Co-Founder & CEO of Daily

​​​Outline:

​We've built a new kind of LLM-powered, multi-user space exploration and trading game as an experiment in what future software might look like. The game has a conversational voice interface and a dynamic UI generated on the fly by an LLM. The core gameplay mechanics are implemented with long-running agents that perform a variety of tasks for users and in the background of the game. In this talk, we'll show how we use sandboxes to make the game flexible and hackable.

​​🕒 6:40 pm - 6:50 pm
Lightning Talk "Building Stateful, Agentic Data Systems for Vibe Analytics"

​​🎤 Lei Tang, Cofounder & CTO of Fabi.ai

​​​Outline:

​Vibe analytics is long-running, iterative, conversational data analysis with AI in the loop. Making this reliable is fundamentally an engineering problem. The core design principle is code-first agents. Agents should take actions and manipulate products the way humans do—by writing and executing code. Agent behavior is therefore compiled into executable code, enabling validation, caching, and reproducibility.

​- Agentic flow and context engineering: Agents operate as constrained collaborators with explicit tools, permissions, and dry-run validation. We explore how to structure analytical context across data sources, intermediate results, and user intent, and how to trade off between RAG, token limits, and prompt caching to optimize correctness and user experience.

​- Stateful, sandboxed execution with smart caching: Analysis runs in isolated, containerized environments with persisted variable state across iterations. Caching is driven by code structure and data dependencies—not chat turns—enabling fast iteration without unsafe re-execution or unnecessary warehouse queries.

​​🕒 6:50 pm - 7:00 pm
Lightning Talk "An Introduction to Prompt Optimization with GEPA"

​​🎤 Sherwood Callaway, Founder & CEO at Stealth Dev Tools Startup

​​​Outline:

​Prompt optimization is usually messy: tweak a sentence, rerun a few examples, repeat. This talk shows a more systematic way to do it. I'll introduce GEPA, a technique for improving prompts using feedback from evals, and walk through how we applied it internally to a complex LLM system without changing models or sharing product details. You'll leave with a practical way to think about prompt optimization as an engineering problem, not a guessing game.

​🕒 7:00 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.