The end of junior developers? How AI is changing IT firms

Junior developers are not disappearing, but their work is changing: the routine tasks juniors used to learn on are now done faster by an AI assistant. Companies therefore need fewer people writing ordinary code and more people who can judge whether code is right, and juniors have to reach that skill by a different route than before.

We are not writing this as forecasters. We are writing it as seniors who work with AI assistants every day and have rebuilt our whole way of working around them. We leave numbers and predictions to others; here is what we see ourselves.

What changed for us

A typical project used to look like a chain: a project manager takes the brief, an analyst writes it up, a senior designs the solution, juniors write most of the code, a tester checks it and the senior glues it all together. Many people, many handovers, many meetings.

Today the same work is done by a senior with AI assistants. They hear the brief directly from the client, AI writes the first draft of the code, tests and documentation, and the senior reads it, corrects it, decides and deploys. A task a junior would once have been given two days for is often done within an hour. Not because AI is smarter than a junior, but because the explaining, waiting and round of fixes are gone.

How the roles are changing

Role	Before	Today with AI
Junior	writes simple code to a spec	verifies, tests and learns to read code they did not write
Senior	designs, reviews juniors, solves hard problems	directs AI, reads every change, stands behind the result
Tester	clicks through the application by hand	decides what needs verifying, writes automated tests with AI
Project manager	relays messages between client and team	drops out on smaller jobs, the client talks to the senior

This does not apply equally everywhere. In large companies with dozens of teams, coordination roles still make sense. For everyday work for small and medium-sized businesses, though, the chain of people is getting much shorter.

What it means for juniors

Honestly: getting into the profession is harder today. Work a company once paid a junior to do so they could learn on it can now be done more cheaply. At the same time, something new opens up: a junior with an AI assistant can learn faster than ever, as long as AI does not do their thinking for them.

If you are starting out, in our experience it pays to focus on this:

• Read code, not just write it. Most work with AI is judging code someone else wrote. Whoever can quickly spot a flaw in a proposal is valuable.
• Understand the fundamentals. Databases, HTTP, security, how a server works. AI replaces typing, not understanding.
• Debug in production. Logs, monitoring, reproducing a bug. AI cannot do this without a person yet.
• Know the domain. Someone who understands accounting, logistics or e-commerce can give AI the right tasks.
• Do not ship anything you do not understand. The worst habit of today is copying an AI solution that works while nobody knows why.

What it means for companies that commission development

For the client the change is welcome: the same money buys more finished work. But watch one thing. Value has moved from writing code to judgement. If a supplier bills you a senior rate but the work is done by a junior with AI and no review, you are paying for speed without a safety net.

So ask who will actually work on the project and who will review every change. Our answer is simple: a senior, at €38 / hour excl. VAT. We describe what that way of working looks like in practice in our AI-powered software development service and on the pricing page.

If you have your own team and are thinking about bringing AI into it, it is not just about tool licences. Review, testing and how new people learn all need adjusting. We help with that through IT consulting.

What an ordinary day for a senior with AI looks like

To make this less theoretical, here is a typical small task. A client writes that they want to change the free-shipping rule on their online store. The senior opens the repository and tells the AI assistant what should change. The assistant finds the related places in the code and proposes a change and a test. The senior reads the proposal, notices the assistant forgot about discount coupons that change the cart total, and has it fix that. Tests pass, CI is green, the change goes to production and the senior checks it on the live site.

The whole thing often takes less time than writing a brief for a junior used to. But notice where the value was added: not in writing code, but in someone knowing that coupons exist and checking for them. That is exactly the skill a junior needs to learn today to earn a place on a team.

Where AI will not replace a person

AI is excellent at writing code that matches the brief. It is weaker with the brief itself. It does not know what the client meant by "simple", it does not know the same thing broke once before, and it carries no responsibility when something breaks on a Friday evening. Architecture, security, talking to the client and deciding what actually gets deployed stay with a person. That is why our rule is that AI is a tool, not an excuse.

A question we do not have a ready answer to

If juniors do not get work, where will seniors come from in ten years? This is a real problem for the whole field and it would be dishonest to pretend we have solved it. We think the path will run through fewer juniors who learn faster and work alongside a senior and AI on real projects from day one. Exactly how that will look, only time will tell.

Summary

AI has not abolished developers. It has abolished a lot of routine and with it some of the roles that were built on routine. For companies that means faster and cheaper development, provided a person with judgement stands behind it. For juniors it means learning to judge sooner, not just to write.

Want to see what senior + AI development looks like on your brief? Send it to us.