If AI does the work juniors used to learn from, who trains the next senior engineers?

Discussions
, , ,
1

Something I keep running into: AI helps experienced engineers more than it helps beginners. Seniors use it to go faster at what they already know. Juniors are supposed to use it to learn — but in practice the two outcomes look very different.

If a senior + AI can now produce what used to take a small team, the “rational” move for a company is to hire the senior and skip the juniors. Fine short-term decision. Bad long-term one — because today’s seniors were yesterday’s juniors who got the reps in: the debugging sessions, the incidents, the code reviews where someone senior explained why an approach would break in six months.

Curious how this is playing out for people here:

  • Are you seeing junior hiring actually shrink where you work, or is it more anecdotal so far?
  • If you’re mentoring juniors right now, has AI changed how you do it — more pairing, different kinds of tasks, anything?
  • Anyone found a way to use AI as a teaching tool for juniors rather than just a shortcut that skips the learning?

I wrote a longer take on this here if useful context: Who Trains the Next Senior Engineers? — QA Redefined — but mostly want to hear how others are actually handling it.

2

AI isn’t closing the software engineering skills gap.

It’s amplifying it.

The best results from AI-assisted coding are showing up with the most experienced engineers.

Not juniors.
Not early-career developers.
Not teams still building fundamentals.
And that creates a real problem.

Because if companies start thinking:
“Why hire more junior engineers when senior engineers + AI can do more?”

…it may look efficient in the short term.

But it weakens the future.
Because experienced engineers don’t just appear.

They are developed.
Over time.
Through repetition.
Through mistakes.
Through feedback.
Through understanding.

That’s the part many teams risk bypassing.
AI can help people ship faster.

But faster output does not always mean deeper capability.

If the tool does too much of the thinking,
learning drops.

And when learning drops,
so does long-term engineering strength.

That’s the real risk with AI in software teams:
More code.
Less understanding.
More speed.
Less judgment.
More delivery.
Less depth..

The companies that win won’t be the ones that just adopt AI fastest.

They’ll be the ones that protect learning while using it.

Because great engineering has never been about output alone.
It’s about people who can think, reason, adapt, and grow.

AI may increase productivity.
But without intentional investment in junior talent, it may also widen the very skills gap the industry can’t afford.

3

I work in SaaS, pre-covid a lot of teams were marketing and pushing the idea of young vibrant energetic teams and junior hires were common and also really good for team culture and balance.

Since covid where a lot of team ended up streamlining and every team member was expected to have strong consultancy skills so decent business acumen and strong communication skills with customers.

This changed the marketing from young vibrant to highly skilled and experienced as the main selling point, customers would not pay the same rates for juniors and as they interview everyone saas allocates to a project they also always picked seniors when they were available.

The market had/has a lot senior availability so recruiting juniors despite the benefits has dropped significantly. We still do university sauna days and other events but recruitment is a smaller number than before though many come back after a couple of years experience and they get picked up then.

Now your enterprise level companies still seem the most obvious route for juniors, I do not have the same insight as to whether they have cut back or not.

The impact of AI is still being worked out.

In some areas if its felt seniors will just resort to what they know well and are not picking up AI they may shift to juniors if colleges and uni’s are preparing them for AI well, we may see a lot of mature folk struggle due to this bias in thought.

Another view is AI can lower entry level to different and get same results so this also opens the door for juniors.

Pairing does make sense but as team sizes drop, it can be viewed as costly.

Personally my view is AI has increased my ability to fast track a bright junior in weeks to solid productivity, not months or years in some companies but I am not working with many juniors at the moment to put that into practice.

I think the market is going to be tough all round due to the uncertainty on how development models will evolve with AI, a lot of companies are wait and see at the moment.

As a junior you would likely want to be networking and find ways to show your personal work, AI allows an individual to do full lifecycle activities so showing real stuff is actually easier for a junior but competition will be high in my view.

4

I don’t think AI is replacing junior engineers as much as it’s replacing some of the experiences that helped them grow.

The best use I’ve seen is treating AI as a mentor not a solution generator. Asking it to explain code, review an approach, or point out edge cases teaches far more than copying an answer.

The challenge is making sure juniors still get enough real-world experience.