We Removed Junior Developers From Our Price List. AI Was Not the Reason

A few weeks ago we took junior developers out of the rate calculator on our website.

It was not a statement about AI. It was bookkeeping. In seven years of running an IT staff augmentation company, clients have asked us for a junior developer maybe twice. Both times were before AI coding tools existed. The line was sitting in our pricing because it looked complete, not because anyone used it.

That is the part missing from most of the conversation about AI and entry-level engineering jobs. In our segment, the demand was already close to zero. AI did not kill it. AI arrived at a funeral that had been going on for years.

What AI did change is the work junior developers used to grow on.

Boilerplate. First versions of tests. Migrations. Reading an unfamiliar module and explaining what it does. Looking something up for an hour before you understand it. That was the ladder. You climbed it by doing unglamorous work until the patterns stopped being mysterious.

Most of that is now a prompt away. Our engineers use Cursor and Claude daily, and the routine part of the job takes a fraction of the time it used to. That is a real gain. It also means the ladder has fewer rungs.

So the question is no longer whether juniors are useful. It is where they are supposed to become seniors.

Our CTO, Alex Sergienko, has a blunt answer to that: the big companies.

His point is that only large firms have the volume and the margin to carry someone through their first year. They can put a junior on a large team, absorb the cost of their learning, and bill around it.

Staff augmentation does not work that way, and it would be dishonest to pretend otherwise. We work contract to contract. A client comes with a specific need and asks for a strong specialist, now. Nobody has ever asked us to send someone who needs teaching first, and there is no reason they would: they are paying for an engineer who contributes in the first week. That is the entire promise. There is no space in that model to train anyone, and no time either.

That is not a moral position. It is arithmetic. But if you multiply that arithmetic across every small and mid-sized firm in the industry, you get a system where almost nobody is paying for the first two years of an engineer's career.

I want to include the argument against my own concern, because our CTO makes it and I do not have a clean answer.

His view: there are enough senior engineers. They are simply busy. And it is not obvious that five years from now the industry will need seniors in the same numbers, or in the same shape, that it needs them today. If AI keeps absorbing the mechanical parts of the job, the definition of "senior" changes with it, and the pipeline question becomes less urgent than it looks.

He may be right. I notice that I cannot disprove it.

What I can say is that the skill nobody has automated is judgment: knowing when the output in front of you is wrong. That skill is not taught in a course. It comes from having done the work badly, been corrected, and understood why. Which is exactly the experience we have stopped paying for.

Nothing dramatic, and I would rather say that plainly than invent a program we do not run.

We do not place juniors, because our clients do not buy them. What we do instead is invest in the people already with us: they stay on the same projects for years, they take on more responsibility inside a team that knows them, and the growth happens on real work with a senior nearby. It is slower than a training program and more honest than one.

It is also the only version of this that survives contact with how we get paid. Nobody is funding a curriculum. What a long engagement funds is time next to someone better than you, on code that matters, with consequences attached. That used to be what a first job was.

For clients, my advice is simpler than it sounds. If you have any in-house team at all, you are one of the few places where an entry-level engineer can still be grown, and the cost of that is lower now than it was, because the routine work they used to be assigned is cheaper to produce. The bottleneck was never their salary. It was senior time. Use AI to buy that time back, and spend it on a person.

And for the senior work that cannot wait two years for someone to grow into it, that is where we come in.

If entry-level work is handled by a tool, and nobody in the middle of the market is paying for beginners, where does the next group of senior engineers come from?

I do not think the answer is "it will sort itself out." I also do not think it is a crisis this year, or next. But it is a question worth asking now, while the people who could answer it are still employed.

That is the part we sell. Our engineers join your team and work on your codebase, inside a process that reviews what gets generated before it reaches your users. 70,000 hours delivered, clients in six countries.

Two or three matched profiles inside 48 hours, monthly terms, and if the fit is wrong in the first week we replace the engineer.

We work mainly in Ruby on Rails, React, Next.js, JavaScript, Java, Golang, iOS, Flutter and Rust.

Talk to us about a senior engineer