AI Reshapes Entry-Level Engineering Hiring in Semiconductors

Graduate recruitment among leading UK employers fell 24% between 2022 and 2026, and AI has entered a labor market already under strain, according to Simon Davidmann, AI and EDA researcher at Southampton University. He cited UK data showing 75% of graduate engineers still find employment within 15 months, a figure he called typical. Davidmann spoke at a Semiconductor Engineering roundtable on college graduates and AI adoption, the first in a three-part series, alongside executives from Cadence, Synopsys, Breker Verification Systems, llmda.ai and ChipAgents.

The panel agreed that demand for electrical engineering graduates has not fallen. What has changed is the work they are expected to do, and the training path that once carried them from junior to senior roles.

What happened to computer science hiring?

Nagesh Gupta, CEO of llmda.ai, said the pain is concentrated in data science rather than computer science. He has been hiring CS graduates from good universities and found few available.

Chuck Alpert, fellow at Cadence, attributed the slowdown to coding assistants. "The vibe-coding tools — Cursor, Claude Code, GitHub Copilot — are improving the productivity of computer scientists, and so it's natural there will be a hiring slowdown," he said. Companies now spend heavily on Claude and OpenAI subscriptions, and they do not want to hire engineers at the same rate.

Thomas Andersen, vice president of AI and ML at Synopsys, split the picture in two. EDA software developers face caution because employers want to see how much AI can automate. Chip designers face a different reality: a long-running talent shortage that AI partially addresses. "Humans are just elevating the tasks that they're doing, and can be more efficient, and then they can develop more chips," he said.

Why are internships disappearing?

Dave Kelf, CEO of Breker Verification Systems, described a cycle that removes practical experience from new graduates. AI is replacing interns, so internships are harder to find. Graduates then arrive with less hands-on skill, and employers must spend a year getting them productive on real designs — a burden that is growing, not shrinking.

He also framed the financial pressure directly. If a company can put AI on junior tasks once used to train new hires, it cannot afford those hires. If it hires them anyway while competitors deploy AI, it loses on profitability.

Davidmann warned of a structural consequence. Without graduates and apprenticeships entering the pipeline, companies never develop the next generation of experts, and AI will not fix the gap left when today's experts retire.

Andersen agreed. "While you can say I don't need the junior guys anymore because AI can do those simple tasks, if I don't replace the existing experts, who's ever going to do it?" he said.

How has mentorship broken down?

Gupta offered a case study from his own family. His daughter, a recent CS graduate, found a job, and nearly her entire graduating class did as well. But she observed that senior engineers now use AI for much of their work and have no time to train juniors. Her large employer has launched a focus group to work out how to make experienced staff take on new graduates again.

Kexun Zhang, head of research at ChipAgents, called the core issue a mismatch. Companies expect AI-native hires who solve higher-level problems with taste and initiative, not the grunt work interns once did. Universities cannot adapt at the pace AI is evolving. "Most of their education has not been AI-native, but the companies that are seeing the economic value of using AI have already changed their expectations for new grads," he said.

Davidmann, drawing on his time running a European university program for a large EDA vendor, proposed splitting engineering education into a foundation mode and a professional mode. The foundation mode covers mathematics, algorithms and debugging without AI. The professional mode treats agents and AI as part of the job. In one project he supervised, students first ran simulations, test generators and coverage manually, then used AI to improve the process, learning both layers.

Can junior engineers teach the seniors?

Alpert said engineers hired within the last two years are now among the best sources of AI expertise inside companies. Three to five years ago, he said, a young engineer could teach him nothing. Today he learns a great deal from them.

Kelf placed the shift in historical context. When logic synthesis arrived, senior engineers working at the gate level resisted it, and new graduates drove HDL adoption. The pattern repeated with randomized verification, where graduates fluent in object-oriented programming picked up test benches faster than senior staff. He suggested AI courses may become intuitive within another year or two, letting new graduates handle issues such as hallucinations that still trip up seniors.

Will electrical engineers escape the disruption?

Alpert said designs have become roughly 10× more complex, driven by packaging, 3D-ICs, multi-field verification and drop testing. Companies want 10× more designs with the same resources and the same timeframes. He argued agents deliver another productivity gain, but the gap will not close. "The number of designs people want to do means there's never going to be a shortage of EEs, even if designers are 10× more productive," he said.

Andersen added that junior versus senior now describes tasks, not tenure. If junior tasks are automated, new hires simply level up from the start, without 20 years of accumulated experience.

Alpert condensed the conclusion into one line: all engineers are now senior engineers. As universities rework curricula and companies rebuild mentorship, the graduates who bridge both worlds — fluent in AI tools and grounded in fundamentals — are likely to command the strongest demand.