Meta Bolsters Superintelligence Team with Key Hires from DeepMind and Scale AI

Meta has intensified its recruitment efforts to build a new superintelligence team, drawing heavily from Google’s DeepMind and the data-labeling startup Scale AI. According to LinkedIn data analyzed by Business Insider, Meta has brought in at least 10 previously unreported DeepMind researchers since July, many of whom worked on some of Google’s most advanced AI systems. Among the key hires is Yuanzhong Xu, a senior engineer at Google DeepMind who played a central role in developing LaMDA and PaLM 2—foundational models behind Google’s Gemini series. Xu’s LinkedIn profile lists his current role at Meta as “TBD,” a likely reference to a unit within Meta’s Superintelligence Labs (MSL) focused on training cutting-edge AI models. Another standout hire is Tong He, whose mathematical prowess helped Google achieve gold medal-level performance at the International Mathematical Olympiad. Meta also recruited three other DeepMind researchers involved in that achievement. Additional hires include Mingyang Zhang, who specialized in improving how large language models retrieve and use information, and Xinyun Chen, whose expertise lies in AI systems capable of generating code and reasoning logically. Meta has also made significant moves into Scale AI’s research division, particularly its Safety, Evaluations, and Alignment Lab (SEAL). At least six Scale AI researchers have recently joined Meta, including Ziwen Han and Nathaniel Li, who co-authored “Humanity’s Last Exam”—a benchmark test designed to push AI models to their cognitive limits. Other new hires from Scale include Alexander Fabbri, Felix Binder, Chen Xing, and Lifeng Jin, all now working on safety and evaluation tasks within MSL. These hires follow Meta’s $14.3 billion acquisition of nearly half of Scale AI in June, which included bringing on Scale’s CEO, Alexandr Wang, to lead the company’s superintelligence initiative. Summer Yue, the original head of SEAL, joined Meta shortly after the deal, now leading the alignment group at MSL. Julian Michael, who succeeded Yue at Scale, also transitioned to Meta around the same time. The recruitment drive underscores Meta’s aggressive push to close the gap with rivals like OpenAI, Google, and Anthropic in the race to develop advanced AI systems. The company is offering competitive packages to lure top talent from across the AI landscape, including former employees from OpenAI and other leading labs. Scale AI’s spokesperson Joe Osborne noted that the movement of employees to Meta is part of a broader transactional shift rather than a reflection of broader industry trends. Meta declined to comment, while Google did not respond to requests for information. As Meta continues to expand its research capabilities, its focus on safety, alignment, and high-performance AI systems signals a strategic effort to not only innovate but also ensure responsible development of future AI technologies.