Unveiling the Mysterious Engineering Teams Behind DeepSeek and Alibaba’s Qwen
The recent surge in interest around DeepSeek and Alibaba’s Qwen artificial intelligence platforms is a topic of keen interest among technology experts and investors in China and the United States. However, the inner workings of the Chinese engineering teams behind these platforms remain largely shrouded in mystery. This analysis leverages publicly available information to shed light on the origins and key contributors of DeepSeek and Qwen, revealing intricate connections to prestigious research institutions and cross-border collaborations.
Separate Paths: DeepSeek and Qwen’s Engineering Teams
Initial research indicates that the engineering teams responsible for DeepSeek and Qwen do not collaborate or share personnel. The only apparent linkage is that DeepSeek’s developers have publicly stated they “distilled” Qwen2.5, along with Meta’s Llama model, to create DeepSeek-R1. This development led to a notable decline in the US stock market in late January.
DeepSeek’s Core Team: Influences and Connections
DeepSeek-R1 has 16 key contributors, several of whom have strong ties to leading institutions. One notable contributor is Yu Wu, who was mentored by Ming Zhou of Microsoft Research Asia (MSRA) during his PhD at Beihang University. Wu also completed multiple internships at MSRA and joined as an associate researcher post-graduation. Another contributor, Daya Guo, was supervised by Zhou during his PhD and mentored by MSRA researchers Nan Duan and Duyu Tang. Zhibin Gou and Zhihong Shao, both from Tsinghua University, collaborated with MSRA’s Nan Duan on research papers.
Zhenda Xie, a contributor to DeepSeek, was advised by MSRA’s Baining Guo during his PhD at Tsinghua University and spent time as a research intern at MSRA. Zhou Ming and Nan Duan, still employed by MSRA, have held leadership roles within the China Computer Federation, including chairing the Computer Terminology Approval Working Committee.
Implications of MSRA’s Role
Though MSRA’s current involvement does not imply a direct stake in DeepSeek, US pressures on China’s tech sector may impact MSRA’s operations. Previously, MSRA halted recruitment from several Chinese universities linked to the military, including Northwestern Polytechnical University (NPU) and Harbin Institute of Technology (HIT).
Alibaba’s DAMO Academy: Founding and Contributions
In contrast, Alibaba’s Qwen was developed by the DAMO Academy, established in 2017 for AI research. DAMO stands for Discovery, Adventure, Momentum, and Outlook. The academy boasts a prestigious advisory board of 10 prominent educators and researchers from both the US and China.
Under the leadership of Jingren Zhou, previously a Microsoft R&D partner and now Chief Technology Officer at Aliyun, DAMO Academy’s data management research has proven pivotal. Chang Zhou, a key engineer for Qwen, joined DAMO in 2017. Zhou, a graduate of Fudan University and Peking University, had prior collaborations with Alibaba’s software engineers.
Academic Background and Impact
Chang Zhou worked with Peking University professor Jun Gao, a prolific researcher with projects funded by China’s 863 Program and the National Natural Science Foundation of China (NSFC). Gao has published over 30 research papers. In 2022, Chang Zhou co-authored a paper with Bin Cui, director of Peking University’s School of Computer Science and deputy director of CCF’s Technical Committee on Databases.
Despite a recent lawsuit filed by Alibaba following his departure to ByteDance, Chang Zhou’s contributions to computer science and his involvement in high-profile projects funded by the 863 Program underscore the prestige of DAMO Academy’s research.
Comparing Qwen and DeepSeek: Strengths and Considerations
Alibaba’s development of Qwen2.5, an original AI model, contrasts with DeepSeek’s reliance on distilled versions of existing models. This distinction likely explains why Apple has partnered with Alibaba to integrate AI features into their iPhones.
From Beijing’s viewpoint, both Qwen and DeepSeek hold strategic importance, contributing to China’s industrial upgrades and mitigating the effects of US tech decoupling and sanctions.
Leadership and Policy Implications
Chinese President Xi Jinping is reportedly set to chair a symposium aimed at boosting confidence in the private sector. Notably, Alibaba co-founder Jack Ma and DeepSeek founder Liang Wenfeng will attend this event.
This gathering underscores the critical role these AI ventures play in China’s tech landscape and their potential influence on future policy and collaboration efforts.
As the technology landscape continues to evolve rapidly, the contributions and partnerships of both DeepSeek and DAMO Academy will likely shape the future of AI development in China and beyond.
We invite you to share your thoughts on these developments in the comments section. Your insights can contribute to a deeper understanding of these exciting advancements in artificial intelligence.
For more updates on AI and technology, subscribe to our newsletter and follow us on social media platforms for real-time news and analysis.