A small school, run by practitioners.
We teach the parts of applied AI that are hard to learn alone: evaluation, deployment, and the judgement to stop a project early.
Neuryn started because a group of engineers kept being asked the same question by colleagues: how do I tell whether this model is actually working. The honest answer took a fortnight to explain properly, and no course covered it.
What we are not
We are not a video library. Nothing here is a recording you watch alone at your own pace with a quiz at the end. Every programme has a cohort, a calendar, an assignment every fortnight and a person who reads what you submit.
We are also not a certification body. The certificate records a brief that two practitioners assessed. It carries the weight of the work behind it, and we say so plainly rather than implying an accreditation we do not hold.
How we are funded
Subscriptions, and nothing else. No vendor sponsors a module, no cloud provider pays for placement in the curriculum, and no tool appears in a programme because someone bought its way in. When we teach a specific product it is because the instructor uses it at work and will say what is wrong with it.
Where we are going
Slowly. Cohorts are capped at thirty because marking quality falls above that, and we have turned down the obvious growth path more than once. The constraint is deliberate: the thing that makes the programmes work is the one that does not scale.
Mara Okonjo
Lead instructor, Evaluation
Machine learning engineer. Builds evaluation harnesses for teams that ship models to production.
Jonas Reiter
Lead instructor, Systems
Systems engineer working on retrieval and serving. Interested in what breaks under real traffic.
Priya Raman
Lead instructor, Governance
Policy researcher turned practitioner. Works on documentation, audit and model accountability.