Run the migration you keep postponing as the training
Train agentic engineering on the legacy migration the team has postponed for years, not on a sandbox. The team learns on its real codebase while the debt shrinks.
Written by
-
Benedikt Stemmildt Founder & Co-CEO
Published: 6 August 2026 Updated: 2 min read
A client asked me last month for a two-day training, and then asked, half joking, whether the team would actually use any of it afterward.
Exercises do not survive the real codebase
Fair question. Most training is exercises. A toy repo, a clean problem, a tidy answer. People learn the moves, go back to the real codebase the day after, and the real codebase looks nothing like the exercise.
So here is the version I now recommend, from our experience. Do not train on a sandbox. Train on the migration you have been postponing.
Every org has a postponed migration
Almost every engineering org has one. The mainframe nobody wants to touch. The Delphi app from two acquisitions ago. The platform with a vendor sunset date that keeps getting closer. It is the biggest piece of technical debt on the books, and it sits there because the work is slow, unglamorous and hard to staff.
That is the wrong reason to keep postponing it, and it happens to make a perfect training vehicle.
How to structure the training
Pick the migration as the training context, not a fake repo. Spend the first morning teaching the agent to read the old system: it generates a plain-language spec of what a legacy module actually does, and the team corrects it. Then the team rewrites that module in the target language with the agent, on the real code, with real tests. Each day takes one more slice. The agent does the reading and the first draft, the engineers do the judgment and the review.
Two returns from the same two days
What the team learns is not abstract. They learn context engineering on the exact codebase they will work in for the next year. And the org gets the migration moving, which is the part finance notices.
The training cost stops being a cost. It buys the team a skill and pays down a chunk of the debt that was blocking the company. Same two days, two returns.
The honest test for any training is whether the work survives contact with your real system. Run it on the real system and the question answers itself.
What is the migration your team keeps postponing, and what would it mean to make that the training?
Further reading
Written by
-
Benedikt Stemmildt · Founder & Co-CEO
More on Adoption
- Find where the work piles up before you pick an AI playbook Four engineering organizations put AI on four different problems. What transfers from them is the diagnostic they all ran first: where does the work actually pile up?
- When the CEO sits in the developer training AI stopped being a tooling purchase and became an investment in the environment: the harness, the context layer, the trained team, the hiring filter. A CEO doing the exercises next to their staff engineers shows the shift.
- Agentic coaching is the new agile coaching Adopting agentic engineering has the same shape as adopting agile twenty years ago: an organizational change, resistance among senior people, and the same three levers that worked the first time.
Scaling needs named owners
Coding agents scale across teams when named people own the platform, the team practice, the guardrails, and the measurement. Rollout & scale is where that ownership is built and handed to internal owners, in your stack.
Training on this topic: Agentic Engineering for Engineering and Product LeadersWant to check one team first? The self-check: twelve yes or no questions, three minutes, instant result, no account. Take the self-check