There is no single rate
The most common mistake in budgeting an AI hire is treating "AI engineer" as one market. It is at least three, and they do not track each other. A number that looks generous for one is uncompetitive for another, which is how a role ends up open for two quarters with an approved budget nobody can spend.
AI Engineer
Applied work: building the product around existing models.
The broadest pool, and the one most adverts are actually aimed at.
AI Research Scientist
Research-leaning: owning whether an approach works at all.
The smallest and most expensive pool of the three.
MLOps Engineer
Platform: serving, cost, observability and the pipelines underneath.
Priced against platform and infrastructure engineering, not research.
What the applied market looks like
The range below is the one most teams are budgeting against, because applied work is what most "AI engineer" adverts describe.
Based on live Artificial Intelligence Jobs adverts for Ai Engineer roles across the UK. Salaries stated in the advert, not offers accepted.
Remote or office-anchored
Genuine remote widens the pool nationally; anchoring to one office narrows it and prices up.
Greenfield or inherited
Maintaining someone else's pipeline is priced as a premium, not a discount.
Someone senior to learn from
Worth more to candidates than most employers think — and it costs nothing.
Read it as what employers are asking the market to accept, not what candidates end up on. Advertised ranges are a negotiating position on both sides, and the top of a band is usually reserved for someone who needs no ramp-up at all.
What actually moves a number
Seniority labels move a figure less than most hiring managers expect, because they are not comparable between companies. These three are, and candidates price all of them:
-
1
Where the work happens
A genuinely remote role competes with every employer in the country and prices accordingly. A role requiring three days in one office is competing for whoever lives near it, and pays for the privilege.
-
2
What state the work is in
Greenfield attracts people cheaply. Inheriting an undocumented pipeline someone else left behind is a premium, because everyone experienced enough to do it has done it once and did not enjoy it.
-
3
Who they will learn from
The single most underpriced thing an employer can offer. A team with someone genuinely senior can hire below the top of the band; a team without one is asking a candidate to accept a career pause.
If your budget is below the range, narrow the role before you raise the number. A well-scoped role at the middle of a band fills faster than a vague one at the top.
Specialisms price separately
The narrower titles are their own markets with their own rates — LLM Engineer, RAG Engineer, Computer Vision Engineer and NLP Engineer among them. If your problem is specifically one of those, budget against that role rather than the general one, and advertise by that name.
Check it against the live market
Every range on this site is drawn from what employers are advertising right now, so the quickest sanity check on a budget is the adverts themselves.