AI coding agents will build exactly what we ask for, even after the system has quietly outgrown the shape that answer used to fit.

This talk teaches agents architectural judgment through a small set of concrete signals that indicate when a growing event-driven system needs something new. It then puts that judgment to the test live, running identical prompts against a baseline agent and a judgment-guided agent as the system grows from a small starting point into a multi-team, multi-repo flow.

We’ll see exactly where the two agents’ code diverge, when the judgment-guided agent stops and thinks, and what architectural signals triggered that decision.

What you’ll learn:

• The signals that tell an agent when an architecture needs to evolve

• A live comparison of a baseline agent and a judgment-guided agent

• How their implementations diverge as the system grows

• What it takes to move AI coding agents from following instructions to making architectural decisions
  
  Agenda
  5:30 – 6:00 PM – Registrations
  6:00 – 7:30 PM – Talk followed with QnA
  7:30 – 8:00 PM – Networking and Refreshments