Which Researchers Could Follow Shengjia Zhao to Meta Next?
Meta’s Super-intelligence Lab has reverberated through the AI community, not least because David Zuckerberg appointed Shengjia Zhao—Stanford PhD candidate and former OpenAI trailblazer behind ChatGPT and GPT-4—as the lab's Chief Scientist. Meta is doubling down on the quest for artificial general intelligence (AGI) and this has provoked speculation in the research community about who might be the next Shengjia Zhao to Meta. In this article, we will consider the names, research directions, and recruiting tactics shaping this new AI "talent war."
Arrival of Shengjia Zhao: Setting the agenda
Shengjia Zhao had only recently made a name for himself as a co-creator of ChatGPT and GPT-4 when he left OpenAI to co-found and lead Meta’s Superintelligence Lab. Meta's decision to hire Zhao was strategic, taking a multibillion-dollar bet on building an elite 'superteam,' together with supposedly competitive product directions, aimed at challenging the of OpenAI and Google DeepMind. With Zhao's unprecedented leadership and scientific aspirations, Meta is now pursuing groundbreaking directions is reasoning models, advancing scaling paradigms, and leveraging multimodal AI systems.
Meta's hiring campaign: Who has been hired already?
Meta is on a hiring spree, and they are looking for talent, even at the highest level, especially if they have experience working together. Over the past few months, Meta has hired at least eleven top researchers from AI labs, and many more appear to be on the way. Notable names who have followed equites like Zhao or have ensured their hire are:
Jason Wei - Worked on the o3 model for OpenAI, and reinforcement learning. He also worked on chain-of-thought prompting, which Zhao studied and has advanced.
Hyung Won Chung - Worked on reasoning and agent-based AI systems at OpenAI, and was the research team lead on decision structuring about AI systems.
Jiahui Yu, Shuchao Bi, Hongyu Ren - All three are former researchers at OpenAI. Yu was co-creator on o3, o4-mini, and GPT-4.1, and Yu is also responsible for multimodal projects (for example, combining text, vision, etc.) at OpenAI.
Trapit Bansal - Another expert in reasoning models from OpenAI.
Three multimodality researchers from OpenAI Zurich, too.
Lucas Beyer, Xiaohua Zhai, Jack Rae, Johan Schalkwy - From other companies, such as Google AI, DeepMind, etc., but all are big names in the area of vision, speech, or foundational models.
This exciting new wave of talent is being led and coordinated by Alexandr Wang, new Chief AI Officer of Meta (previously Scale AI) and Zhao, who are collectively establishing a new frontier AI vision for the company.
Why Are Researchers Transitioning to Meta?
The reasons for the attractiveness of Meta to researchers are consistent and strong:
Vision and Infrastructure - The vision of AGI, access to large compute clusters (such as the 1-gigawatt Prometheus facility that will be operational by 2026), and the freedom to work on next-generation model architectures, are incredibly enticing.
The Money - Good reports of $100M-$300M total compensation packages (with stock vesting immediately with leadership roles) has further compounded Meta's ability to attract the best and the brightest in AI.
Synergy and Scale - Meta focuses on putting teams together to achieve synergy, often hiring researchers with a history of working together to enhance communication and speed.
Who Could Follow Zhao Next?
Meta has established a precedent: it not only hires researchers with Zhao's interests in aspects of reasoning, scaling, and multimodal AI, especially those who could help him foster a synergistic vision, but also hires into those capabilities.
The candidates most likely to be next in line are:
Researchers Already Writing with Zhao: Those who published papers with Zhao for either OpenAI or Stanford, or those who worked on chain-of-thought and multimodal AI are also targets.
Star Researchers in Reasoning and Alignment: Those with expertise in model alignment, safety, and interpret-ability may want to join Meta's growing super-alignment, especially as Meta will advance efforts around safe super-intelligence.
AI Scientists Frustrated at Competitors: OpenAI and Google DeepMind researchers who would value open-sourcing their work (as Meta has pledged) may follow Zhao to Meta, or prefer to work for Meta at its scale and leadership.
Top Areas of Interest to Keep an Eye On
Meta’s next hiring priorities (and possible new colleagues joining Zhao):
AI Reasoning & Chain-of-Thought Research: Looking at the o1 and o3 models.
Multi-modal Models: Vision, speech, text, and cross-modal reasoning—building on the next GPT-4o-like systems.
Post-training and Alignment: Safe and transparent AGI based on OpenAI’s alignement teams.
Synthetic Data Creation: Innovation for scalable model training, a Zhao specialty.
Autonomous Agents: Long-horizon, action-taking AIs are increasingly driving Meta’s product aspirations—so agent-based researchers could be next.
The Conclusion
Meta is transforming into a global AI powerhouse under Zhao's view by aggressively hiring leading researchers and fostering a culture of depth in innovation and collaboration. If Meta continues to build momentum, we can expect to see more and more people join Zhao's team, including alumni from OpenAI, DeepMind, and even FAIR (Meta's original AI organization), who will follow Zhao's leadership towards super-intelligence. The new "talent war," means that the best researchers in AGI, reasoning and alignment will soon be heading to Menlo Park.The next wave of AI innovations will happen in teams following Zhao's lead, which announced an ambitious forward vision, backed by a mountain of computing power, with everything else that comes with its incentives; and similarly in business as well, there is power in following a visionary leader through the transformative experience.
Stay tuned for a while. The Met-AI's club of geniuses has merely started to set up the towers.
Post a Comment