AI training
Your team using AI on your own work, not on a course's examples.
A course built on somebody else's repository teaches your team to use a tool on somebody else's problem. We build the material around your stack, your codebase and the work already on your board — in English or Spanish, on site or remote.
What you get
What your team can do the week after.
Every format starts with a discovery call and a look at how your team works today. The examples are yours, and so is what follows.
- Tell what these models do and do not do, precisely enough to plan with
- Write a specification an agent can execute, instead of a long chat nobody can review
- Run a plan-execute-review workflow, and recognise the point where each part breaks
- Put guardrails around agent work: tests, boundaries, and checks that could actually fail
- Review work nobody on the team wrote, at a volume nobody used to see
- Choose a model on cost and latency rather than on the last thing they read
- Keep company data out of a model, and spot a prompt-injection attempt
- Measure whether any of it helped, on your own numbers
How it goes
Four steps, and the first one is about you.
Discovery call
Half an hour on what your team builds, what they already use, and what would count as this having worked.
before anything
Material built on your work
We write the exercises against your repository and the tickets already on your board, so nobody spends the day on a toy project.
the week before
The sessions
On site or remote, in English or Spanish, with everybody typing rather than watching. Half a day, two days, or a shape you asked for.
your calendar
What you keep
The material, the exercises, the working setup on everyone's machine, and a short written note on what to do next.
yours afterwards
Which one is you
Four formats.
Executive briefing
Half a day, for the people who decide. What is real, what is theatre, what it costs to run, and what changes about hiring and about risk. No live coding.
half day · leadership
AI in the daily workflow
Two hands-on days for a whole team. Where these tools help in the work your team actually does, where they quietly cost more than they save, and how to tell the difference this month rather than next year.
2 days · any team
Agentic engineering for developers
For engineers who already use an assistant and want the next thing: agents working from written specifications, a harness that constrains them, checks that catch them, and reviews that scale past one reader.
2 days · engineering
Custom curriculum
Built from a discovery call when none of the three fit — a specific stack, a regulated environment, a team spread across three time zones.
scoped per team
Questions
What you will want to ask.
- How many people can attend?
- Up to about a dozen for the hands-on formats, because everybody types; the executive briefing scales further.
- Can you run it in Spanish?
- Yes, and in English, including rooms where the material is in one language and the questions arrive in the other.
- What do we need to have ready?
- A repository we can look at, one real piece of work in progress, and machines your team is allowed to install things on.
- Do people leave with a certificate?
- No — they leave with the material, a working setup and exercises done on your own code, which is the part anybody would actually check.
AI training · next step
Tell us about the team.
Size, stack, language, and what they are already using. That is enough for a proposal.