Cover Image for No Evals, No Optimization, No Production.
Cover Image for No Evals, No Optimization, No Production.
Avatar for MEGA Code
Presented by
MEGA Code
Registration
Approval Required
Your registration is subject to host approval.
Welcome! To join the event, please register below.
About Event

AI agents are moving from demos into production. But once they start touching real workflows, teams face a new set of questions:

Are they doing what we expect?
Where do they break?
How do we trace failures back to the right step?
And once we find the problem, how do we actually improve the agent?

This meetup is for builders who want to move beyond “it seems to work” toward measurable, debuggable, and continuously improving AI agent systems.

What We’ll Cover

  • How to evaluate AI agents beyond simple pass/fail checks

  • Why observability matters once agents move into production

  • How runtime traces help teams understand where agent behavior starts to break

  • How to connect agent failures back to workflow steps and code context

  • How evaluation and debugging signals can become optimization inputs

  • How AI agent systems can improve through iterative optimization loop

What to expect:

  • 5:30 - 6:00: Arrival, Pizza, Drinks, Networking

  • 6:00 - 6:20: Talk – Runtime debugging for agents, with step-level reproduction, handed off to your agent via MCP.

  • 6:20 - 6:30: Demo – Finding the failing traces, understanding what went wrong, and patching the code, all through MCP.

  • 6:30 - 6:50: Talk - How agent optimization moves teams beyond debugging individual failures

  • 6:50 - 7:00: Demo - MEGA Workbench turning eval signals into iterative performance gains.

  • 7:00 - 8:00: Q&A, then Networking

Pizza and beverages will be provided.

Save your spot, spaces are limited.

Venue graciously provided by HAC.

About the Hosts

MEGA Code

MEGA Code builds self-evolving AI agent optimization infrastructure, focused on evaluation-driven development, reusable agent wisdom, and optimization loops that help AI agent systems improve across runs.

Elastic Dash

ElasticDash is a step-level debugging tool for AI workflows, helping developers identify and resolve issues at each stage of their agent and LLM pipelines.

HAC

Hanwha AI Center (HAC) is a private membership for all those dedicated to AI. As a hub for innovation, HAC brings together passionate entrepreneurs, researchers, and visionaries to explore the societal and technological impacts of AI on human life.

Location
Hanwha AI Center
300 Grant Ave Suite 500, San Francisco, CA 94108, USA
Avatar for MEGA Code
Presented by
MEGA Code