There was a time when having a long list of skills on your CV felt like an obvious advantage in tech. If you could work across frontend and backend, understood databases, knew your way around the cloud and could pick up whatever framework the team was using, you were useful in more places.

That logic has not disappeared. In fact, plenty of companies still want people who can move comfortably across a product. What has changed is what employers seem to expect underneath that versatility.

A recent analysis from Revelio Labs looked at 75 million tech job postings and found that since 2024, the average number of skills requested in a role has fallen by roughly 25%, from 30 to 21. At the same time, the experience expected in those skills has increased.

Employers are asking for fewer things, but they appear to want people who know those things better.

So, is the generalist actually losing ground?

The answer is more complicated than it first appears.

When Knowing a Little About Everything Was Enough

The appeal of the generalist is easy to understand, particularly in smaller engineering teams. If you have ten developers rather than a hundred, having someone who can move from an API issue to a frontend problem without declaring either one “not my job” is genuinely valuable.

That has not suddenly become a bad thing.

What may be losing value is a particular version of the generalist: someone whose breadth comes mainly from familiarity.

They have worked with React, Python, AWS, Docker, PostgreSQL and perhaps a dozen other technologies. They can get started in many places, but when a problem becomes difficult, unfamiliar or consequential, somebody else has to take over.

AI makes this distinction much more important.

Today, getting started with an unfamiliar technology is considerably easier. A developer can use an AI coding assistant to explain a codebase, generate a first implementation, translate syntax, suggest tests or walk through a framework they have never used before.

In other words, some of the advantage that came from simply knowing your way around many technologies is becoming easier for other people to acquire.

The harder advantage is knowing enough to recognise when the obvious solution is wrong.

That requires depth.

Perhaps We Have Been Thinking About Specialists the Wrong Way

Specialisation often sounds restrictive. You choose one lane, become exceptionally good at it and leave everything outside that lane to somebody else.

Modern software rarely works that neatly.

A backend engineer still makes better decisions when they understand how the frontend consumes their APIs. A cloud engineer benefits from understanding application architecture. A security engineer who understands how developers actually build and deploy software is likely to be far more useful than one who treats security as a completely separate function.

Interestingly, current hiring data reflects some of this messiness.

Research released by Andela on September 10 analysed 47,101 technical job postings from Fortune 500 companies and found that 53% of the AI and machine-learning roles it examined required skills associated with at least two established roles. The research identified emerging combinations such as MLOps Pipeline Engineer, LLM Application Engineer, FinOps Reliability Engineer and DevSecOps Security Engineer.

These are hardly narrow jobs.

They are specialised jobs built from multiple disciplines.

That distinction matters because the future may not belong to the person who knows only one thing any more than it belongs to the person who knows a little about everything. It may belong to people who have developed serious expertise somewhere and enough range around it to apply that expertise in different situations.

Think of it less as choosing between depth and breadth and more as deciding where your depth begins.

Job Titles Are Starting to Struggle With This

There is another interesting signal in Andela's research.

The company identified 23 recurring combinations of technical skills that did not map neatly onto standard job titles. It also found 6,758 postings containing the skill combination it classifies as “LLM Application Engineer,” even though employers were not necessarily calling the jobs that.

That tells us something about the speed at which technical work is changing.

We still talk about careers using familiar labels such as frontend developer, backend developer, data scientist and DevOps engineer because labels make hiring easier. The actual work increasingly crosses those boundaries.

This is where declaring the death of the generalist becomes difficult.

Another 2026 dataset, from EngRadar's State of Tech Hiring 2026, found full-stack engineering to be the largest category in the vacancies it tracks, with 11,790 open roles as of July 31. The report itself cautions that job titles are imperfect measures because backend and frontend work can sit inside jobs labelled “full-stack.” Still, it is useful evidence that companies have not suddenly stopped valuing people who can work across a system.

So we have two things happening at once.

Companies still need range. They are also becoming more particular about expertise.

The Real Question Is Where You Are Deep

This changes the career conversation.

Tech professionals have spent years being encouraged to keep adding skills. Learn another framework. Add cloud. Understand DevOps. Learn AI. Pick up data. Keep moving because nobody wants to become obsolete.

There is nothing wrong with learning widely. Curiosity is one of the things that makes good technical people good.

The problem begins when collecting technologies becomes a substitute for developing expertise.

Twenty technologies on a CV can look impressive, but they tell an employer very little about the kinds of problems you can actually be trusted to solve. Someone with deep backend experience, for example, may understand databases, infrastructure, security and frontend development too. The difference is that when the conversation becomes difficult, there is somewhere in their skill set where familiarity turns into judgment.

That judgment is becoming particularly important when AI can produce plausible answers so quickly.

The valuable engineer is increasingly not the person who can produce an answer first, but the person who understands the problem well enough to know whether that answer should make it into production.

So, Is the Generalist Losing Ground?

Maybe the better answer is that the shallow generalist is.

The evidence is not telling us that companies suddenly want developers who live inside one tiny technical box. Full-stack roles remain significant, while newer jobs are combining capabilities that previously belonged to different disciplines.

What appears to be changing is the value of breadth on its own.

Knowing a little about many things was more distinctive when acquiring that knowledge took considerable time. AI has made access to information, syntax and basic implementation much cheaper. What it has not made cheap is the experience behind good technical judgment.

The strongest generalist in this market may therefore look quite different from the old jack-of-all-trades.

They can still cross boundaries. They can still learn quickly and speak the language of several parts of the business. They simply have something underneath all that range: an area where they have gone far enough to understand not only how things work, but why they fail, which trade-offs matter and when the easy answer is not good enough.

So perhaps the question for someone building a career in tech is no longer how many things they can add to their skill set.

It is what they are willing to know deeply enough that their judgment becomes valuable.