§ 01 · The Landscape

“The UK tech market is worth £1.2 trillion. The interesting question is who actually benefits from that.”

Imran, thank you for joining us. Let’s start with the macro picture. What is the state of the UK tech labour market heading into 2026?

It is two markets layered on top of each other, and people who confuse the two end up making bad career decisions. The first market is the headline one. The UK tech sector is worth around £1.2 trillion and remains the largest tech ecosystem in Europe. London accounts for about 59 percent of that value. Venture funding is flowing back after the 2023 trough, and AI specifically is taking close to 20 percent of all UK venture investment. That is the market the press writes about. The second market — the one that actually matters to people reading this — is the hiring market underneath. That market is much tighter, much more selective, and much less generous than the headline numbers suggest. Companies are spending more on AI tooling and less on incremental headcount. The result is a smaller number of senior, well-compensated roles competing fiercely against each other, sitting on top of a much smaller-than-it-used-to-be junior pipeline. If you are in the senior layer, 2026 is good. If you are trying to enter the field, 2026 is hard. Both things are true simultaneously.

The LSE work we drew on for the briefing identifies ten in-demand roles. Are those the right ten, in your view?

Broadly yes, with one caveat. The LSE list captures the categories where employer demand is genuine and where compensation is rising. AI engineers, data scientists, data governance specialists, AI product managers, cybersecurity professionals, cloud and DevOps engineers, full-stack developers, UX designers, and site reliability engineers — all of those are legitimate categories with real hiring activity. The caveat is that the headline category names hide enormous internal variance. “AI engineer” covers people doing fundamentally different jobs at fundamentally different compensation levels. A research engineer at a London foundation model lab is a different role than an MLOps engineer at a fintech is a different role than someone calling themselves an AI engineer because they wire OpenAI APIs into existing software. All three appear under the same label in the surveys. Read the categories carefully. The label is the start of the analysis, not the end.

And the salary numbers? Are the figures we cite accurate, in your view?

The averages are roughly right. What the averages hide is the bimodal distribution. UK tech compensation is splitting more sharply than the means suggest. The senior tier — experienced specialists in London, in particular sectors — pulls salaries dramatically higher than the published averages. Often double, sometimes triple. The entry tier is being pushed downward as AI tooling compresses what employers will pay for junior work. The mean number masks the fact that there are two markets, and the gap between them is widening every year. The professionals doing well in 2026 are well above the published averages. The ones doing poorly are well below. The average increasingly describes nobody specifically.

§ 02 · Reading The Table

“The AI effect column is the part most career-switchers should read first.”

Walk us through the table. What is the most important thing for a reader to notice?

The right-hand column. The AI effect column is the part most career-switchers should read first, because it tells you which roles AI tooling is making more valuable, which it is making less valuable, and which sit in the uncomfortable middle where the answer depends on how you specialise within the role. Read the salary numbers, but read them through that filter. A senior AI engineer earning £200,000 plus is a different bet than a full-stack developer earning the same number, because the underlying trajectory of the two roles is going in opposite directions. The senior AI engineer’s compensation is being supported by an AI hiring boom. The senior full-stack developer’s compensation is being supported by legacy demand that is starting to compress.

Which roles in the table get the strongest tailwind from AI?

Three categories, clearly. AI and machine learning engineering, AI product management, and data governance and AI ethics. Those three roles benefit directly from every additional pound that flows into AI tooling. Demand is racing ahead of supply, particularly for senior practitioners. The senior tier in those roles is genuinely well-paid and likely to stay that way through 2026 and 2027. Cybersecurity belongs in the augment column for slightly different reasons — AI is creating new attack surfaces faster than it is creating defensive automation, which means more cyber threats, which means more demand for skilled cyber professionals. Demand has been outpacing supply for a decade in cyber, and AI is widening the gap, not closing it.

And the disrupted roles?

Full-stack development is the clearest disruption story in the table. The entry-level demand for general-purpose full-stack developers has measurably softened over the past eighteen months because AI tooling has taken over much of what junior full-stack developers used to do — CRUD endpoints, basic frontend scaffolding, boilerplate API integration. Companies that used to hire two juniors and a mid-level now hire one mid-level with AI tools. The role is not disappearing, but the entry pipeline into it is narrowing fast. The senior end remains paid well because senior full-stack developers do architectural and integration work that AI does not handle well. But the middle is being squeezed.

§ 03 · The Hidden Categories

“Data governance is the role nobody is talking about, and it is the cleanest entry path right now.”

Of the ten roles, which is the most underrated in your view?

Data governance and AI ethics. The role barely existed five years ago. Now it sits at the centre of every AI deployment that any regulated organisation undertakes, and the supply of qualified practitioners is nowhere near demand. The reason it is underrated is that it sits at the intersection of three skill sets that rarely overlap in one person — technical literacy, regulatory understanding, and organisational judgement. People who can credibly hold all three are scarce, and salaries are rising fast as a result. The category is also welcoming to people coming from non-engineering backgrounds. Lawyers, compliance professionals, policy researchers, social scientists, even philosophers — all of them can move into this work with the right additional technical grounding. It is one of the few tech-adjacent roles where a humanities background is genuinely an asset rather than a hurdle.

What about AI product management? You flagged that as having acute shortage in the table.

Yes, and worth being specific about what is driving it. Product management as a discipline has existed for decades. AI product management has been a real category for maybe two years. The number of people who genuinely understand both what AI models can do and how to translate that into shippable, monetisable product is small, even at large tech companies. Demand is enormous. By 2026, more than three-quarters of UK product leaders are planning to expand AI investment, which means they are hiring AI PMs faster than they can find them. The compensation reflects that. Senior AI PMs in London are routinely earning above £150,000 base before equity, and the role has unusually fast progression paths because the talent pool is so thin.

Cybersecurity has been “hot” for a decade. Is the demand really still increasing?

It is accelerating, not just continuing. The category started the 2020s with a chronic skills shortage and has not solved it. The current AI wave is making the gap worse, not better, because AI is enabling much more sophisticated attacks. Cybersecurity has effective zero percent unemployment among qualified practitioners. The senior compensation in cyber is unusual — truly exceptional specialists can earn well into £250,000 plus for specific niches like cryptography, offensive security research, or financial-sector incident response. The entry barriers are real, though. You cannot bluff your way through a cyber interview the way you might in some adjacent tech roles. The technical depth required is substantial, and the responsibility is significant.

§ 04 · Entry Points

“If you are switching careers in 2026, the most efficient path is not the one your friends took in 2018.”

Take a reader who is currently in marketing, design, or a non-technical professional role. They want to move into tech. Where do you actually point them based on the table?

Honestly, in three specific directions. First, data analytics. The barrier to entry is reasonable, the demand is consistent across nearly every industry, and someone coming from a non-technical role with strong domain expertise has a genuine competitive advantage over a fresh graduate. A marketer who learns SQL, Python basics, and Tableau brings something a CS graduate does not have — understanding of the business context the data is meant to inform. Second, UX and product design, particularly the AI-adjacent end of product design. The shortage of designers who understand how AI-native interfaces work is acute. Anyone with design taste, user empathy, and a willingness to learn the technical patterns of AI products has a clear path. Third, the new category — data governance, AI ethics, AI policy work. The path requires building technical literacy alongside whatever existing professional expertise the person has, and the destination is a role with both compensation and meaningful societal impact.

What about formal education? Is a Master’s degree or certificate worth it for someone switching careers?

Depends entirely on the role. For data governance, AI ethics, and AI product management, structured education from a credible institution genuinely helps. These roles are partly about signalling — you are being trusted with regulatory, ethical, or strategic decisions, and employers want some external validation that you understand the territory. Specialised programmes from institutions like LSE, the various Russell Group universities, and a small number of credible private providers carry real weight in hiring decisions for these roles. For pure technical roles like full-stack development or DevOps engineering, formal credentials matter less. What matters is the portfolio — what you have shipped, what you can demonstrate. The pathway depends on which destination you are aiming at, and being honest about that difference saves a lot of wasted effort.

The LSE work you cited identifies a 54 percent figure for firms struggling to fill entry-level digital roles. If demand is that strong, why are entry-level salaries not rising faster?

Because the gap is not really about quantity. It is about the mismatch between what entry-level candidates can do on day one and what employers need them to do on day one. Companies report difficulty filling entry-level roles, but they are not lowering their standards to fill them — they are leaving the roles open or using AI tooling to do the same work. The candidates being hired at entry level in 2026 are unusually capable. They have portfolios. They have shipped things. They have demonstrated AI fluency. They are not the same population as the bootcamp graduates of 2018 who could walk into a junior role on the strength of a CS-style curriculum. The bar has risen, and the salary has not risen with it, because the supply at the new bar is roughly matching demand at the new bar. That is the dynamic underneath the numbers.

§ 05 · Geography

“London is still where the senior compensation lives. The regional gap has not closed.”

London accounts for 59 percent of UK tech sector value. How much of that translates into a salary premium?

For senior roles, substantial — typically 20 to 30 percent above national averages, with concentration in particular postcodes. The senior AI engineer earning £200,000 plus is almost certainly working in London or remotely for a London-headquartered firm. For entry-level and mid-level roles, the premium is narrower — perhaps 10 to 20 percent — and gets partially eaten by cost of living. The honest answer for someone weighing London versus regional UK is: if you are early in your career, the regional cities are increasingly competitive. Manchester, Bristol, Edinburgh, and Cambridge all have strong tech ecosystems with meaningful senior opportunities, and the cost-of-living differential makes regional roles materially more attractive than the headline numbers suggest. If you are aiming at the top of any of the ten categories in the table, London is still where the highest-paying roles cluster, and remote work has not changed that as much as the 2021-era predictions suggested it would.

Has remote work meaningfully redistributed UK tech jobs?

Partially. Mid-level roles have spread out considerably — full-stack developers, data analysts, DevOps engineers can credibly live anywhere in the UK and work for London firms. Senior roles have proven stickier. The senior layer still gravitates toward London because the deal flow, the conference circuit, the casual lunch meetings that lead to the next role — all of that infrastructure still concentrates in central London. Remote work is real, and meaningful, but the prediction that London would lose its tech-talent gravity has not played out the way it was forecast. The city remains the centre. It just has a wider catchment area now.

§ 06 · Closing

“Pick the role where you have an unfair advantage. The salary table does not show you that part.”

Last question. If a reader is looking at the ten roles and trying to choose one to pursue, what is the single most useful piece of advice you can give them?

Pick the role where you have an unfair advantage that nobody else in the candidate pool has. The salary table tells you which roles pay well. The salary table does not tell you which role you specifically should pursue. The answer to that depends on what you bring that is not on the standardised list of skills. A former lawyer moving into AI governance has an enormous unfair advantage. A former teacher moving into UX research has an enormous unfair advantage. A former finance analyst moving into data science in fintech has an enormous unfair advantage. The trap people fall into is picking the role with the highest headline salary and then competing against people who have ten years of relevant background you do not have. The role you should pick is the one where your existing background lets you skip three years of competition. Identify what that is for you, specifically, and start there. The numbers in the table are interesting context. They are not the answer. The answer is somewhere in the intersection of what the market wants and what only you can credibly offer. Find that intersection and the path becomes obvious.

Imran, thank you.

Thank you. The next conversation is the more useful one — readers writing in to say where their unfair advantage actually sits. That is when this turns from analysis into action.