AI titles have multiplied faster than the work has changed. We now see “AI engineer”, “AI product lead”, “head of AI” and “AI transformation director” attached to roles that would have been called something else two years ago.
Some of that is real. Some of it is title inflation.
The practical question is not whether a candidate has an AI title. It is whether they have done the work that the role actually requires.
Separate the work from the label
In our searches, applied AI roles tend to fall into four groups:
- Research and model development. Deep technical work, often with a research background.
- Applied engineering. Building products and systems that use models, with a strong software engineering base.
- Platform and operations. Making models usable, reliable and affordable in production.
- Product and commercial. Deciding where AI creates value and how to measure it.
A candidate can be excellent in one group and a poor fit for another. The title rarely tells you which.
Look for evidence of shipping
For applied roles, the most useful signal is evidence of something shipped, operated and measured. That might be a model in production, a data pipeline, a product feature or an internal tool. The candidate should be able to explain the trade-offs: cost, latency, quality, risk and maintenance.
A research background is valuable, but it is not a substitute for product judgement. A strong software engineering background is valuable, but it is not a substitute for understanding model behaviour.
Assess the foundations
We assess applied AI candidates on the same foundations as other senior engineers: system design, data handling, testing, reliability and communication. Then we add the AI-specific dimensions: model selection, evaluation, failure modes and cost.
That combination is harder to find than a candidate with an AI title. It is also more useful.
Be honest about the market
In our searches, senior applied AI engineers across our European markets typically sit in the £95,000–£130,000 range, with a premium for people who have operated models at scale. Remote European roles benchmark lower. Research scientists with a strong publication record can command more, but the premium is not universal.
The market is not as irrational as it sometimes looks. Companies pay for evidence of delivery, not for the word “AI” on a CV.
Look at adjacent backgrounds
Some of the best applied AI hires come from adjacent engineering backgrounds: backend engineers who have worked with data-intensive systems, platform engineers who have run GPU workloads, product engineers who have built evaluation tools.
They may not have the title. They often have the foundations. A good assessment process will find them.
The test
Ask three questions:
- What have they built that uses a model in production?
- How did they evaluate whether it worked?
- What would they do differently next time?
The answers will tell you more than the title ever will.