Author :
Diana Sharma
We have more than 300 industry partners across New Zealand and Australia, and a decent chunk of my week is spent talking to them about what is and is not working in junior hiring.
Those conversations have changed noticeably in the last two years. I want to set out what they now say, partly because our partners deserve to hear their own feedback reflected back, and partly because if you are considering a move into tech it is a much more accurate picture than any job advertisement.
Nobody has ever told me they lost a junior hire over the wrong framework.
Not once. Every hiring manager I talk to accepts that a junior arriving from any training pathway will need to learn their stack, their conventions and their domain. That is assumed. It is a few weeks of work and everybody has priced it in.
Job ads still list technologies, because job ads have to list something and because that is how the applicant tracking system filters. Treat that list as a rough indication of the environment, not as the actual bar.
The actual bar is somewhere else entirely.
Four things come up over and over, in almost the same words, from organisations that have nothing to do with each other.
They cannot debug outside a tutorial. The person can follow a known path competently. Then they hit an error nobody has written a Stack Overflow answer for, and there is no method. No forming a hypothesis, no narrowing the search, no reading the error properly. This is the single most common complaint and it is not a knowledge gap. It is a practice gap.
They take the AI's suggestion without judging it. This one is new and it is now the fastest-growing item on the list. The tools are genuinely good and everyone is using them, our own teams included. The problem is a junior who cannot tell the difference between a suggestion that is right and one that is confidently wrong, and who therefore cannot be trusted with anything that has not been reviewed line by line. The value of a junior developer has shifted towards exactly the thing AI does not supply, which is judgement about the output.
No system-level thinking. They change one thing and something unrelated breaks, because they never formed a mental model of how the pieces connect. They were taught components. They were never asked to hold a whole system in their head.
They cannot reason out loud. Asked why they made a decision, they cannot reconstruct it. This is the one that quietly ends the most placements, because a person who cannot explain their thinking is extremely hard to help. Mentoring them is guesswork.
Those four are the same failure in different clothing: learning that took place entirely in a clean environment with a known right answer at the end. Real work has neither.
It is worth being straight about the structural change underneath this, because it is not the fault of anyone entering the industry now.
The old model assumed the employer did this part. You hired a graduate, absorbed six to twelve months of low productivity, and taught them how work actually works. That was a reasonable deal when teams were bigger and the pace was slower.
Teams are leaner now, they ship faster, and that runway has largely gone. Meanwhile the routine work that used to be a junior's training ground, the small well-defined tasks, is precisely the work that automates most easily.
So the entry-level job has not disappeared. It has changed shape. It now starts at roughly the point it used to reach after six months, and employers need people who arrive already able to operate, not people who will be ready by Christmas.
That is uncomfortable and I am not going to dress it up. It is also the entire reason our programmes are built around mentored work with real industry partners rather than around exercises. If the workplace can no longer provide the first six months of practice, the training has to.
When I ask what the good ones had in common, the answers are consistent and not very glamorous.
They had been critiqued before. Someone had reviewed their work properly and they had learned to take that without either collapsing or arguing.
They asked early. The team could see them thinking, and could help. Several partners have described this as the single best predictor.
They had finished something. Not a large thing. Something complete, that other people used, where they had to handle the boring last twenty percent.
They could hold a conversation about a trade-off. Even a simple one. Why this approach and not that one, what it costs, what they would do with more time.
And, for the career changers specifically, they brought the prior career with them. Partners tell us this constantly about people who came from nursing, teaching, hospitality, the trades. They already know how to handle an angry customer, run a handover, work a shift where things go wrong. Those are senior behaviours in a junior role and they are visible in the first fortnight.
Two thoughts, from watching several hundred of these placements.
Hire for the trajectory, not the current state. The gap between a good junior at month one and month six is enormous, and it correlates much more with how they respond to feedback than with what they knew on arrival.
And be honest with yourself about what your team can offer. The placements that go badly are almost always ones where nobody had time to supervise. A junior with no reviewer is not cheap labour, they are an unreviewed risk. If you can give one person two hours a week to look at their work properly, this model works extremely well. If you cannot, wait until you can.
1300 graduates, and four out of five job seekers are in work within twelve months. That number is not a training outcome. It is what happens when employers tell us honestly what is not working and we build the programme around the answer.
If you would like to talk about hosting a placement, or about what your team is finding in junior hiring, get in touch. Those conversations are where our programmes come from.
Own what's next.