09TrainingAI Training
Capability you keep, not a subscription.
Your team doesn’t need another tool demo. They need to tell when a model is the right answer, how to check whether its output is any good, and where the risk actually sits. We teach that on your systems, your data, and your workflows — so it survives contact with Monday morning.
- Shape
- Role-based cohort, run over weeks
- Format
- Live sessions on your own code and documents
- You leave with
- Judgment, not vocabulary
01 — The problem
Enthusiasm without judgment
Half the company is pasting confidential text into whatever tool they found last week. The other half refuses to touch any of it. Both positions come from the same gap: nobody has been taught how to evaluate these systems.
- Output accepted because it reads well, not because anyone checked it.
- Contract text and customer data sent to tools nobody approved.
- Vendor claims nobody can verify, so the demo makes the decision.
- Engineers rebuilding what a library already does — or trusting one they never read.
02 — What you get
Training built on your actual work
Every session runs on your documents, your codebase, and your workflows. Nothing is hypothetical. People leave with working artefacts and one shared standard for what ‘good enough to ship’ means — including the cases where the answer is no.
- A curriculum split by role — engineering, product, operations, leadership
- Live sessions where each exercise runs on your own material
- An evaluation harness for one real workflow, with pass criteria your team wrote
- A written data-handling standard: what may be sent where, and to which tools
- An internal playbook of approved tools, approved uses, and hard boundaries
- Recordings, exercises, and worked solutions your team keeps for new joiners
03 — How it runs
Assess
A short skills and usage review — including the tools already in use unofficially — so the programme starts where your team actually is.
Teach
Live sessions per role. Each one ends with something built on your own material rather than a case study.
Apply
Supported project work: the team ships one real improvement using what they learned, reviewed by the people who taught it.
Anchor
Playbook, evaluation criteria, and materials handed over so the capability outlasts the cohort.
04 — Proof
20+
Years of practitioner experience
1000+
GitHub stars on our open source
PhDs
And industry veterans on the team
The people teaching are the people building. The same engineers who deliver Fortune 500 client work — including StoneX Group — run the sessions, so questions get answered from production experience rather than from a slide. Our open-source packages, RedisOplog and @bluelibs/nova among them, are used by developers worldwide.
05 — Straight answers
Our team is non-technical.
Then their track is judgment, risk, and workflow rather than code. We run separate tracks so nobody sits through the wrong session.
Can’t they just learn this online?
Some will. Most won’t. And no free course can tell them whether an approach fits your data, your contracts, or your regulator.
How long until we see a return?
The applied phase ships a real improvement inside the programme itself. That is deliberate — training that ends at the certificate rarely changes behaviour.
The tools will have changed in six months.
They will. We teach evaluation and risk, which transfers. Tool specifics are the smallest part of the curriculum, and the part we expect to date badly.
Buy a tool and you rent capability. Teach the team and you own it.