Build an AI that checks and fixes its own work: agentic loops in practice

Most AI presentations stop at "wow", a nice-looking output that falls apart the moment it meets a real requirement.

In this workshop I'll show you how to go beyond the demo and build an agentic system that works in a loop: one that takes a design from Figma, plans the implementation, writes the code, then validates, fixes and publishes its own work, all the way to a finished website live on the public internet.