The AI Talent War: Tech Giants Are Recruiting PhD Students Two Years Before They Graduate

Hu Qi, a PhD student at the University of Hong Kong, specializes in AI agent safety and is about two years away from defending his dissertation. Headhunters are already contacting him with concrete offers and asking for his graduation date. His story reflects a growing trend: tech giants no longer wait for students to finish their studies.

How recruiters get to candidates before graduation

Hu Qi is interning at one of China's tech giants. His team leads regularly ask about his plans after the defense and insist he move to a full-time position immediately after getting his degree, with no gap.

At the same time, headhunters are calling to pitch specific roles and want to know exactly when he will finish the program. All of this is happening roughly two years before the candidate actually enters the job market.

• Hu Qi's specialization is AI agent safety, one of the scarcest profiles of 2025–2026
• Place of study: doctoral program at the University of Hong Kong (HKU)
• Time until defense: about 2 years
• Offers are coming simultaneously from headhunters and directly from the company where he is interning

Why can't companies wait for graduation?

The shortage of AI specialists in niche areas, including the safety of agentic systems, is so severe that a "let's wait until they finish" strategy has become too risky. While one company waits, another has already made an offer and signed a contract.

Doctoral programs take four to six years, and the industry's demand for talent is growing faster. Corporations are adapting by moving the recruiting funnel as early as possible and locking in a candidate's interest at the very start of their studies, before they have even begun thinking about the job market.

"This reflects a growing sense of urgency," writes the

South China Morning Post, describing the hunt for academic AI specialists.

What PhD students gain and what they lose

For researchers with the right specialization, the situation creates a rare negotiating position: they can choose an employer years before graduating, outside competition with the open market. They can negotiate an offer as a scarce resource rather than as just another job seeker.

But the opportunity comes with pressure: companies want an answer now, not in two years. A PhD student has to choose between locking in terms early and risking a wait for a more competitive field of offers after the defense. AI agent safety has moved sharply to the forefront with the spread of autonomous systems, and specialists in this area are still few and far between.

What this means

The war for AI talent has moved into university lecture halls: companies are building recruiting funnels years ahead, not out of caution but because of a real shortage. For the industry, this is a signal that the AI talent shortage is not a short-term spike but a structural problem.