Agents can produce convincing backend features fast. The hard part is reviewing it and deciding whether the generated code actually solves the problem and if you want to maintain this code in the future.

In this session, we'll do exactly that and use an agent as an active collaborator that writes Kotlin. We'll go into engineering a harness for Kotlin development covering skills I find useful, LSP & IDE support for your agent, quality gates the code needs to pass - both deterministic and probabilistic - as well as agent-friendly tooling and try to provide a heuristic to decide when you should actually read & understand the code.
The goal is not to build perfect prompts or hands-off delivery, but to build a repeatable way to move faster while keeping design, code quality and ownership where it counts.
Leave with an understanding why and how agents thrive when writing Kotlin and how to reliably steer them so that the code fits your teams' requirements and standards.