Everyone has watched the demos. Far fewer teams have shipped anything. We teach the part in between — on Claude Code, ChatGPT, Grok and open-weight models — in workshops built around your codebase, your data and your syllabus.
Each track is taught against a real repository, not a slide deck. Start with the tool your team already uses, or take two and compare them honestly.
The same material, delivered at the intensity that fits. Every format ends with something running that you built yourself.
One day · 6 hours live
One track, one day, one working artefact. We spend the morning on the mental model and the afternoon building, and nobody leaves without something running on their own machine.
Two days · 12 hours live
Day one covers the ground the intensive covers. Day two applies it to your codebase, your data and your constraints, with time set aside for the problems that only appear in your environment.
Four weeks · two evenings a week
Spaced practice beats a single long day when you want the material to stick. Two live evenings a week, a project that grows across the four weeks, and a cohort channel that stays active between sessions.
Scoped to your syllabus
Built around what your course actually requires. Send us the module outline, assessment criteria or dissertation brief and we shape the sessions and the final project to match, at student rates.
One to five days · on-site or remote
A curriculum written after we've read your code and talked to your engineers. Your repository, your data, your deployment target — and the awkward internal constraints that generic training always skips.
Practical briefs, reviewed by a working engineer and discussed live. No multiple-choice paper, and every credential records the model generation it was assessed against.
Training gets a team ready. Sometimes you also need the first system in production. We run those engagements alongside your engineers, and hand it over so it doesn't leave with us.
The tools your team keeps asking for and nobody has time to build — document handling, reporting and the spreadsheet that has quietly become critical infrastructure.
Agents scoped to a job with a clear boundary, a clear escalation path and enough tracing that you can explain any individual decision afterwards.
Nothing useful lives in isolation. We connect the system to your existing services through MCP or plain APIs, with authentication and permissions handled properly.
We build alongside your engineers and train them as we go. The engagement ends with your team owning the system, the tests and the runbook.
Start with the self-paced catalogue. Many of the fundamentals are free, and workshop participants get the matching courses included with their seat.
For the Claude Code, Open Source Models and Agentic Systems tracks, yes — you should be comfortable in a terminal. The ChatGPT and Grok tracks have a route through for technical product people who can read code but don't write it daily.
That's the most common request. Send us your syllabus, module outline or a description of your stack, and we write the sessions and the final project around it. Students and university groups get their own rates.
Live, always — the value is in the questions and the debugging. Cohort sessions are recorded for participants, but the workshops are deliberately small enough that everyone gets attention.
Each one is a practical brief with a live defence. You build something, a working engineer reviews it, and you talk it through. A credential you can't explain isn't one we'll issue.
The tool-specific material is revised every cycle and the certifications record the model generation they were assessed against. The durable parts — tool design, context management, evaluation — change far more slowly, which is why they're weighted heavily.
Yes. Build engagements run discovery, prototype, production and handover, with your engineers pairing throughout so the system doesn't leave with us.
Send us the stack, the syllabus or the problem. We'll come back with a format, a curriculum outline and a date — usually within two working days.