About the Talk #

Remember when deploying to Kubernetes meant hand-crafting YAML manifests line by line? Then build tools started generating them for us. Then AI came along and for a while, that just meant copying and pasting from ChatGPT.

We’ve come a long way. Today, you can deploy pods, troubleshoot clusters, and manage infrastructure through a natural conversation with no YAML required. But what does that actually mean for the people running Kubernetes day to day?

In this talk, I’ll trace the evolution of how we interact with Kubernetes from manual manifests to AI agents that understand your cluster. I’ll demo the open-source Kubernetes MCP Server and explain how the Model Context Protocol is becoming the bridge between AI tools and infrastructure. And I’ll look ahead to where this is going: autonomous agents that investigate incidents and propose fixes before you’ve even realized something was wrong.

This isn’t a talk about replacing platform engineers. It’s about how their role is evolving and why understanding your systems deeply matters more than ever.