AI strategist, AI engineer, ML engineer: what is the actual difference?
6 minute read. Updated 2026-08-08.
An AI strategist assesses whether AI can solve your problem and sets the roadmap; that title barely exists in the advertised market, so expect to price it near broad AI rates. An AI engineer integrates existing models like LLMs into your product, priced close to the £550 Generative AI and LLM median. An ML engineer builds and deploys custom models, priced higher at £575 a day or £76,000 in salary. Hire for the symptom, not the label on the CV.
3 labels, one confused market
AI strategist, AI engineer and ML engineer get used interchangeably in job ads, LinkedIn headlines and agency pitches. They are not the same job, they don't cost the same, and hiring the wrong one for your symptom wastes a quarter before anybody notices.
None of the confusion is really the specialists' fault. Titles in this market move fast, and a single person often does 2 of these jobs across a career without ever changing their CV heading. The fix is to stop reading titles and start matching the role to what's actually broken.
AI strategist: the role the market barely prices
An AI strategist tells you whether AI is worth pursuing for a given problem, and if so, how: build, buy, or wait. It's advisory work, usually short, usually senior, and it precedes any engineering. You need one when the honest question in the room is 'should we even do this', not 'how do we build it'.
This title has no clean row in the advertised market. Nothing at all is published for AI strategy, AI governance or AI risk roles as distinct job titles, the same way nothing is published for AI product roles. The closest proxy is the broad Artificial Intelligence category, which shows a median day rate of £551 and a median permanent salary of £71,000, and even that figure blends every job that gets called 'AI' rather than reading strategy specifically.
Treat that absence as information, not an oversight. Strategy and governance work in AI is largely being bought under borrowed titles right now, so expect to negotiate this one on judgement and reputation rather than a published benchmark.
AI engineer: builds with models that already exist
An AI engineer, as the term is mostly used today, builds products on top of existing foundation models: retrieval, prompting, agent orchestration, wiring an LLM into a workflow your team already runs. You need one once the strategy question has been answered and you're past 'should we' into 'let's build it with what's already out there'.
This work is priced closest to the Generative AI, Large Language Model and Prompt Engineering rows. Generative AI shows a median day rate of £550. Large Language Model and Prompt Engineering roles show the same £550 median, though the market is too thin there to publish a spread. Treat £550 a day as your working anchor for this kind of build.
ML engineer: builds and runs custom models
An ML engineer trains and deploys models from your own data, rather than calling somebody else's. You need one when the problem doesn't fit a general-purpose model, or when the model has to sit inside a production pipeline with monitoring, retraining and rollback built in.
This is priced higher than generic AI engineering work. Machine Learning Engineer shows a contract median of £575 a day and a permanent median of £76,000. Where the work tips into keeping models running reliably at scale, look at the MLOps row instead: median day rate £575, but a permanent median of £87,500, the highest permanent figure in this set. That gap tells you where the market thinks the hardest, least glamorous work sits.
The 2 roles either side: data scientist and data engineer
Neither of these is an AI engineer or strategist, and both get pulled into AI hires by mistake. A data scientist explores your data and tests whether a pattern exists worth building on; median day rate £600, the highest of the specific roles in this table. A data engineer builds and maintains the pipelines that feed everything above it; median day rate £500, the lowest of the specific technical roles here, but nothing above it works without it.
If your actual problem is 'our data is scattered across 3 systems and nobody trusts it', hiring an ML engineer to fix that is expensive and wrong. Hire the data engineer first.
Match the symptom to the hire
'We don't know if AI is worth pursuing here' is a strategist problem, priced near the broad AI median because no cleaner figure exists. 'We want to bolt an assistant onto our existing product using an off-the-shelf model' is an AI engineer problem, priced around the £550 Generative AI and LLM median. 'We have our own data and need a model trained and put into production' is an ML engineer problem, priced at a £575 median day rate or £76,000 median salary. 'Our models exist but nobody owns keeping them alive' is an MLOps problem, the best-paid permanent role in this table at £87,500 median.
The single most expensive mistake in this list is hiring an ML engineer to answer a strategist's question. You end up paying production-engineering rates for someone to spend 3 months telling you what a senior generalist could have told you in 3 weeks.
What to do about it
- Match the hire to the symptom, not the job title on the CV.
- Expect to negotiate AI strategy and governance work without a published rate benchmark.
- Anchor AI engineering work that builds on existing models to the £550 Generative AI and LLM median.
- Anchor custom model building to the Machine Learning Engineer median of £575 a day or £76,000 salary.
- Do not hire an ML engineer to answer a strategy question.
- Fix the data pipeline first if the problem is trust in the data, not the model.
Questions people also ask
Is an AI engineer the same as an ML engineer?
No. An AI engineer, as the term is mostly used now, builds on top of existing models such as an LLM accessed through an API. An ML engineer trains and deploys custom models from your own data and owns the production pipeline around them. The market prices them differently too: AI engineering work sits close to the £550 Generative AI and LLM median, while Machine Learning Engineer work carries a median day rate of £575 and a median permanent salary of £76,000.
Why is there no rate for an AI strategist?
Because it isn't tracked as a distinct job title in the advertised UK market. Nothing is published for AI strategy, governance or risk roles specifically. The closest available figure is the broad Artificial Intelligence category, at a median day rate of £551 and a median salary of £71,000, and that blends every role labelled 'AI' rather than isolating strategy work on its own, so treat it as a rough anchor rather than a clean benchmark.
We already have a data scientist. Do we still need an ML engineer?
Probably, if you want a model in production rather than in a notebook. A data scientist's job is to find out whether a pattern in your data is worth building on. An ML engineer's job is to build and run the thing that uses it reliably, day after day, at whatever scale your business needs. They are complementary roles, not substitutes, and the market prices them slightly differently: £600 median day rate for data science against £575 for ML engineering.
How do I avoid hiring the wrong one of these?
Write down the actual symptom before you write the job title. If the honest sentence is 'we don't know whether this is worth doing', hire a strategist. If it's 'we want to add AI to something that already works', hire an AI engineer. If it's 'we have data nobody has built on yet', hire a data scientist or an ML engineer depending on how far along the work already is.
Where the figures come from
Every rate and salary quoted in this article is a median or percentile of figures advertised in UK job postings over the six months to 8 August 2026. They are not rates paid, and the gap widens at the top of a range.
- IT Jobs Watch, UK contract rates, 6 months to 8 August 2026, read 2026-08-08.
- IT Jobs Watch, UK permanent salaries, 6 months to 8 August 2026, read 2026-08-08.