Building AI-based tooling yourself has become easier than ever. And doing so doesn't mean you should rebuild a hyperscaler's AI platform. Instead, it means keeping control of your own tools, reviewing guidelines, integrations, where it runs and what models get used.

This live-coding session will explore building your own tools, using a code-review bot as an example. We'll build a command-line Spring AI application using Kotlin, providing the model with custom tools and MCPs along the way and get the agent to produce structured output so that review findings can drive deterministic workflows.
Then we'll switch gears and turn the agent into a Kubernetes workload and automating the scheduling of reviews by using a Kubernetes operator. We will tour how to define custom resources and how to build a controller that creates supporting resources, creates a job and delivers continuous status updates.

You will leave with an understanding how to integrate AI into your applications & create your own tools using Spring, and how to build custom cloud automation to run these workloads in your own properly isolated environments.