Agentic Coding for Engineering Teams

Coding agents can speed a team up — and just as effectively generate tech debt that nobody reviewed. The difference is not the tool; it is how you work with it: the context, the task boundaries and the review gates. This workshop teaches you to use agents so that the gain in speed does not come at the cost of the architecture.

Who it is for

A workshop for engineering teams that want a real speed-up from AI without losing control of quality and architecture. It works both where agents are only now entering the team and where they are already in use without shared rules and the effects are showing up in reviews.

How we run it

We work on a real repository and team workflow, not a toy example — short introductions interleaved with exercises, focused on guardrails, not just speed. You leave with concrete working rules you can apply straight away.

What you take away

• A method for choosing the mode of work — assistant, agent, autonomous — for the kind of task and its risk
• The ability to give an agent context, instructions and boundaries so the output is predictable
• Review gates and a test strategy for generated code
• The ability to recognise common traps and tech debt from AI code before they reach production
• A way to measure real impact — productivity and quality — instead of relying on the feeling of speed