For the senior engineer whose review queue has become cleanup duty, and for the leader who bought the seats and still can't see the capability.
Thirty minutes. No deck, no script. We look at where you actually are with AI-assisted development, where you're trying to get, and whether the program is a fit. If it isn't, you'll leave with a straight answer and a few ideas anyway.
You'll be talking to Troels, the person who wrote the training and reviews the work, not a sales rep.
Most people on this call come from one of two places. Both are welcome; the conversation adapts to the seat you're in.
Most AI-engineering courses teach thousands of students per cohort. This one takes five, because the work gets reviewed, not just watched.
The short questions before the calendar help me prepare, so answer them briefly and honestly. You'll get a calendar invite with the video link straight away. Need a slot that isn't listed? Email [email protected].
Thirty minutes is someone's attention, yours and mine. Here's an honest read for each seat.
Not ready for a call? Start with the free guide at start.troels.im, it's the same method, applied to one repo in an afternoon. One honest filter: if you're expecting me to promise 10× productivity, I genuinely don't want us to have that call. I won't make that promise, and we'd both be wasting thirty minutes.
Four beats. Same arc the call always follows (pain, picture, gap, path), read from whichever seat you're in.
For engineers: compounding errors: plausible output, missing context, edge cases, hidden assumptions that quietly become your problem.
For leaders: the review burden and hidden risk behind the seat licenses, demos, and "AI productivity" anecdotes.
For engineers: who you want to be as code generation gets cheap: judgment, specs, verification, recovery, repo readiness.
For leaders: what visible AI capability looks like: clearer specs, sharper review, fewer production surprises, a story you can tell without theatre.
For engineers: session-by-session habits vs. a repeatable system you could hand to a teammate, and why scattered blog posts don't close it.
For leaders: private experimentation vs. an operating discipline you can govern, review, and explain, and why tool rollout alone never gets there.
For engineers: whether the program fits your stack and goals, which cohort start works (September, October, or November), and what it costs to keep things as they are.
For leaders: team seats, scope, timing, and who else needs to be aligned before this becomes real.
Background in quantum computing research, compiler development, and blockchain infrastructure. Principal engineer who has spent the last several years shipping production software with AI as the primary tool, including the CompoundCoders platform itself, built almost entirely with AI assistance.
Talks with engineers about repository structure, context engineering, evaluation, and recovery. Talks with CTOs and CEOs about turning private AI usage into a teachable, governable capability.
The program is a distillation of what actually works in production: the systems that built real products, not theory. The call is a chance to ask the person who wrote it whether it's right for you, directly, with nothing in the way.
Thirty minutes, a straight conversation, and a clear next step, whether that's a cohort seat, team seats, or just a better plan than you arrived with.