Eight roles. Observable work.

Start with the work your team needs to do. Each guide describes four proposed levels, an example work sample and evidence to discuss.

These are proposed role profiles, not measured European proficiency levels. Use relevant work evidence and a consistent assessment process. A country score says nothing about an individual.

Lead & adopt

AI product lead

Choose the problem, define success and own the trade-offs.

Build & run

Applied AI engineer

Build AI features that work beyond the demo.

Build & run

Machine learning engineer

Train, evaluate and maintain models for a defined task.

Build & run

Data engineer for AI

Make data usable, traceable and available to the right people.

Build & run

MLOps / AI platform engineer

Keep AI services deployable, observable and recoverable.

Govern & evaluate

AI governance & assurance lead

Turn risk decisions into checks people can carry out.

Lead & adopt

AI adoption & learning lead

Help teams change their work and measure the result.

Research

AI research scientist

Create and test new methods with reproducible evidence.

Four levels of responsibility

Foundation means demonstrating a bounded task. Pilot means testing it with supervision. Production means owning reliable work in use. Organisational means setting practices across teams. The levels describe scope and evidence, not years of experience or a job title.

One person may cover several profiles, especially in a small team. Separate responsibilities before deciding whether you need another hire.