What's happening with early-career software hiring right now?

The bottom of the ladder is being pulled up. techUK reports that Adzuna has seen a 30% drop in UK entry-level job postings since ChatGPT launched, with graduates facing the toughest market since 2018. The picture inside the sector is sharper still: The Register, citing the Institute of Student Employers, says UK tech cut graduate jobs by 46% in the past year, with a further 53% drop projected.

Engineering specifically shows a widening gap. A GOV.UK snapshot of entry-level hiring puts entry-level engineering hiring at -19% against -10% for senior roles, a nine-percentage-point gap. And the wider market is subdued: ONS vacancy figures show total vacancies down to 712,000 for April to June 2026, with professional, scientific and technical activities among the largest fallers.

So why keep hiring juniors at all? Because AI has changed what a junior does, not whether you need one. The question is no longer whether someone can write a function from scratch. It's whether they can tell when the machine got it wrong.

Which early-career engineering skills are in demand?

Demand hasn't vanished, it's concentrated. Computer Weekly's tech recruitment outlook describes a cautious 2025 with clear hotspots for specialist skills in AI, data, enterprise applications and cyber security. The Register, reporting Accenture data, found demand for AI skills up nearly 200% in a year, with London accounting for 80% of AI-related postings.

That demand is being met through new routes, not just graduate schemes. The DSIT AI Labour Market Survey shows apprenticeships have risen from 3% of AI hires in 2020 to 19% in 2025. Meanwhile cyber security keeps its own pressure: GOV.UK data counts around 143,000 people in the UK cyber workforce, and the AI Labour Market Survey 2025 notes 70% of cyber firms reported at least one hard-to-fill vacancy.

For engineering roles, most of this talent is home-grown. A GOV.UK review of IT and engineering professionals finds the vast majority of new engineering hires come from the resident labour force. In other words, the people who will verify your AI output are already here. You just need to test for the right thing.

How do you hire early-career engineers who verify what AI writes?

Start by rewriting what 'good' looks like. The skill you're buying is critical review: can someone read AI-generated code, spot the plausible-but-wrong answer, trace an edge case and explain why it fails? That's testable. Give candidates a working sample of AI output that contains a subtle bug or a security gap, and ask them to talk you through it.

• Set a code-review task, not a blank-page task. Hand over AI-written code and ask what they'd change and why.
• Plant a realistic flaw: an off-by-one error, an unhandled null, a hard-coded secret. Watch whether they catch it.
• Ask for their reasoning out loud. You're hiring the questions they ask, not just the answer they land on.
• Test how they check their own work: tests written, assumptions stated, sources verified.
• Value curiosity and scepticism over speed. The engineer who says 'let me confirm that' is the one you want.

Look beyond the graduate pipeline, too. With apprenticeships now a real share of AI hires, and a Low Pay Commission report describing a 'low hire, low fire' market, the people who can do this work may not arrive with a computer science degree. Structure your interview around evidence of verification behaviour, and you'll widen the pool without lowering the bar.

How can we help you hire early-career engineers?

This is where we come in. We search 15 million candidates to find early-career engineers who bring the judgement your team needs, and we manage the process end to end: sourcing, screening, shortlisting and booking interviews. Our recruitment agent manages recruitment end to end for 8% on a successful hire, with no monthly fee and no upfront cost, through Reed.ai. Tell us what good looks like, and we'll surface people who can verify what AI writes.