How to seize the opportunity: what a junior can do and what a company must do
What happens to a profession when the first rung everyone climbed starts to disappear?
A few weeks ago I wrote on LinkedIn that at Orbitant we’ve barely hired any juniors; the reactions made it clear the issue goes well beyond our own hiring. Junior developers in the AI era have opened up a conversation that starts with the tasks that used to fall to that role and are now handled by AI, but which quickly reaches the future of an entire industry: who will replace today’s seniors when they retire?
That thread led us to open a roundtable joined by two more people who have spent decades building teams: Jordi Miró (a physicist by training and Zynap CTO) and Víctor González (Fractional CTO).
What makes someone senior, and why AI doesn’t shorten the road
To understand what makes someone senior, it helps to first establish why someone is considered junior. Jordi defined it by the lack of demonstrated experience and by the famous ten thousand hours of flight time, though he added an important nuance: “it isn’t about age, because you can retrain at forty and still lack that mileage, which is exactly what separates a senior from a junior”.
Víctor approached it from the other side, with an image that stuck with me: if you put a senior in a helicopter, drop them into the middle of the jungle, and leave them alone, they’ll survive and make it back to camp. And if they’re one of the good ones, they’ll show up wearing the tiger’s skin.
A senior is someone you leave alone in the middle of the jungle and they find their way back. A junior needs someone walking beside them.
VÍCTOR PACHECO
Fractional CTO
Naming that difference was decisive: the miles matter more than the years — how many problems you’ve worked through, how many things you’ve broken and fixed. That’s what epistemology calls “tacit knowledge”, the kind of knowing that doesn’t show up in books and gets built by wrestling with clients, with hard problems, and with seniors who correct you.
That’s the crux, because AI handles explicit knowledge beautifully while tacit knowledge resists it — and tacit knowledge is exactly what separates a senior from someone who only writes code. What makes someone senior is the judgment accumulated over years, and judgment isn’t something you can download.
Junior developers in the AI era: why the market has narrowed the way in
The data confirms that entry door is closing. The study Canaries in the Coal Mine, by Erik Brynjolfsson and his team at Stanford documented a 13% drop in employment for workers aged 22 to 25 in the occupations most exposed to AI, software development included, with a revealing nuance: jobs whose knowledge is applicable from day one are falling sharply, while those that depend on accumulated judgment keep growing.
AI is absorbing precisely the part of the work a junior could do without experience — which was the door people used to walk through into the trade. Víctor drove the point home with a figure: in June 2026, at one of the most advanced AI companies, barely a handful of engineers had less than three years of experience; almost all of them were PhDs with a track record in another domain.
Jordi separated two worlds that are worth keeping apart. On one side, the large, stable company that hires so many people every year that it maintains its talent pipeline almost by obligation; that’s the case with Netflix, which for a long time only signed seniors until it understood that feeding the culture requires young talent, as its Chief Product and Technology Officer, Elizabeth Stone, has explained. On the other hand, the startup plays a different game, one where stability is nowhere to be found, and you have to run just to keep from falling behind.
Jordi explains it with the new-driver analogy: you start out in an old, cheap car, precisely because a dent barely matters; only after piling up miles and confidence do you earn the wheel of a Formula 1 car, which forgives nothing. A startup, which needs to move fast to survive, is exactly that Formula 1 car; that’s why a small team looks for very senior people, with command of the business and the ability to take ownership, hands already trained for that pace. A junior fits there with difficulty: the context doesn’t give them the margin they need to cut their teeth. The risk shows up in the medium term, because if nobody trains juniors on account of the short-term slowdown, in a few years there will be no seniors left to hire.
Follow our journey building AI agents from the inside
Fundamentals don’t expire, even when the tooling changes
If the market is narrowing the way in, the question becomes what’s actually worth learning. I came in with a doubt of my own: for years we trained juniors in design patterns and refactoring techniques, so I wondered whether that still makes sense, or whether it was like optimizing memory while the world was already moving from assembly to C.
Without knowing the fundamentals well it’s very hard to sustain the critical thinking the job requires: someone has to be able to say “this is going to production” while understanding what sits underneath, Jordi insisted. In his view, you can’t think about complex systems without having built simple ones first, because that’s where you learn to anticipate what can break. Víctor added an interesting nuance: what matters today is understanding how your system scales, how it behaves under load, and what business sense it makes.
With AI, we’ve moved up another level of abstraction: you can now ask the tool about architecture and stress-test its answers, which only adds value if you bring judgment to the conversation. Out of that came the distinction Víctor repeated most, and it’s a key one for anyone starting today:
- Vibe coding is for exploring and validating a quick idea; for that it works beautifully.
- Engineering with agents plays in another league: taking it to production with real users and scaling from a hundred to a hundred thousand, where the problem is the same even though the engineering challenges look nothing alike.
I’ve lived it in my own team: in six months, we’ve gone from talking about prompt engineering to the context engineering system I built to amplify my work and half a dozen more labels, but it all comes down to the same thing — engineering applied to new tools.
The best team is never simply the sum of the best individuals.
JORDI MIRÓ
CTO Zynap
How to seize the opportunity: what a junior can do and what a company must do
If fundamentals matter but speed pushes them aside, what’s left to solve is how to train people without slowing down. Víctor has spent more than a decade with a formula that predates today’s AI and still works for him. He summed it up in a few steps that are useful whether you lead a team or have just joined one:
- Define what to explore, those areas of value the core team can’t tackle today.
- Put the juniors there, away from the critical path, on side projects rather than the product the business runs on.
- Always walk alongside them, because exploring means giving slack with a safety rope, nothing like abandoning someone in the jungle.
- Integrate whatever works, accepting that many threads will die while others come out ready for the real product.
He illustrated it with a case of his own: an intern spent a few months, as a side project, exploring a tool to auto-label their datasets; when the internship ended and they left, the team integrated that work into the product, and it worked very well. And he stressed two things: that this costs intention more than money, and that any company building a product can do it if it accepts a moral contract with the person it brings on board — give them experience and prepare them for the real world.
When asked what they’d look for when hiring, Jordi summed it up in one word: hunger, that attitude that’s hard to measure but easy to recognize in an interview, the one belonging to someone who admits they barely know anything and is burning to learn. If you’re a junior, that’s your best lever, because today’s speed also makes it possible to learn faster than ever with a system of AI skills.
Both agreed on something that puts the rest in order: a company can’t be led by juniors, because seniors set the direction while juniors contribute and grow. And as Jordi reminded us with the example of the last World Cup, the best team is never the sum of the best individuals.
That’s why the decision about today’s juniors is a decision about the company you’ll be tomorrow: whoever stops training people in order to move fast will end up depending on a senior market that keeps getting more expensive and scarcer, while whoever builds their talent pipeline now will have, five years from now, the judgment money can’t buy.
Here’s the video of the full session ⤵️
