Hiring is selective, not slow. Against a subdued 2025 economic backdrop, businesses took a cautious approach but demand held up for specialist skills in AI, data, enterprise applications and cyber security, according to Computer Weekly's tech recruitment outlook for 2026. The wider picture reads as low hire, low fire: recruitment is weaker and vacancies have dropped below pre-pandemic levels, as the Low Pay Commission's 2025 report describes it.
Yet the sector is not standing still. The largest volume increase in vacancies from August to October 2025 came in professional, scientific and technical activities, up by 5,000, per the ONS Vacancies and jobs bulletin for November 2025. Demand for AI skills, meanwhile, rose nearly 200% in a single year, with London taking 80% of AI-related postings, according to Accenture data reported by The Register. When AI features ship at scale, the running cost of every prompt and the speed of every reply become budget lines. That is why briefs now name them.
The tilt is towards proven, specialist engineers rather than volume graduate intakes. UK tech cut graduate jobs by 46% in the past year, with a further 53% drop projected, according to figures from the Institute of Student Employers reported by The Register. Entry-level postings have fallen sharply too: Adzuna recorded a 30% drop in UK entry-level postings since ChatGPT launched, as techUK reports. That squeeze on juniors raises the bar for the skills teams actually screen against.
Two skill clusters stand out. Cyber remains hard to staff: 70% of cyber firms reported at least one hard-to-fill vacancy, per the AI Labour Market Survey 2025, while the UK cyber workforce sits at roughly 143,000 people and is growing again, according to the government's Cyber security skills in the UK labour market 2025. And routes into AI work are broadening: apprenticeships rose from 3% of AI hires in 2020 to 19% in 2025, per DSIT's AI Labour Market Survey 2025. Token cost and latency literacy is fast becoming part of that specialist screen.
Start by writing the constraint as a real requirement, not a buzzword. Ask candidates how they would cut inference cost per request or bring a p95 response time under a target, and what trade-offs they would accept. This works because most new hires in engineering come from the resident labour force, with low reliance on international recruitment, as the government's Professionals in IT and engineering review notes. The talent is close to home; your brief just has to describe the work precisely.
Then widen the funnel rather than narrow it. With entry-level engineering hiring at -19% against senior at -10%, a nine-point gap, per the government's snapshot of entry-level hiring, teams that only chase senior names compete for a thin pool. Apprenticeships and adjacent-skill hires can carry cost-and-latency thinking if you assess for reasoning, not job titles. Move quickly, too: labour-market conditions shift month to month, as tracked in the KPMG and REC UK Report on Jobs, so a slow process loses the people who can actually reason about inference economics.
When the brief is this specific, matching matters. We search 15 million candidates, rank a shortlist in under 30 seconds, contact matched people in under a minute and can book an interview in under three minutes, so you spend your time assessing how someone thinks about token cost and latency, not chasing CVs. Our recruitment agent manages recruitment end to end for 8% on a successful hire, with no monthly fee and no upfront cost. Post your software role on Reed.ai today and see your shortlist.
