On the afternoon of March 3, 2026, Alibaba's Tongyi Lab Qwen team underwent a significant personnel shake‑up. Qwen's technical lead Justin Lin (林俊旸) formally submitted his resignation and in the early hours of March 4 posted a brief farewell on the X platform: "me stepping down. bye my beloved qwen." The news quickly exploded across the global AI community, drawing intense attention from developers, investors and industry observers.
According to reports by LatePost, 36Kr, Reuters and TechCrunch, Lin's resignation was not an isolated event. On the same day, post‑training lead Yu Bowen (郁博文) also left; core contributors to Qwen3.5, VL and Coder, including Kaixin Li (李凯鑫), announced departures on X; and Binyuan Hui (惠彬原), who left in January 2026 after leading Qwen Code, had previously been a key team member. The Qwen core technical ranks therefore suffered an unusually large exodus. This once high‑profile team—whose models have been downloaded more than 600 million times and whose open‑source influence is second only to Meta's Llama—now faces the most serious organizational crisis since its founding.
Justin Lin: From Peking University to Alibaba's Youngest P10
Born in 1993, Justin Lin graduated from Peking University with a bachelor's degree in computer science and a master's in linguistics and applied linguistics from the School of Foreign Languages. He joined Alibaba Damo Academy's Intelligent Computing Lab via campus recruitment in 2019 as an algorithm engineer, became Qwen's technical lead in 2022, and was promoted in 2025 to Alibaba's youngest P10‑level technical executive.
Lin led Qwen from 0 to 1: in 2024 he drove the full‑size open‑source release of Qwen (0.5B–72B), building a large open ecosystem; in 2025 he launched the trillion‑parameter flagship Qwen3‑Max and formed an embodied intelligence team aiming to take models "beyond the screen." Lin was not only the technical core but also Qwen's external face—frequently sharing progress on X and at events like AGI‑Next; even Elon Musk had liked some of his posts.
Resignation Process: Organizational Restructuring as the Trigger
Sources close to Alibaba say Lin's resignation was directly related to ongoing organizational restructuring of the Qwen team.
The Qwen team originally used a vertically integrated model, with Lin overseeing the full chain—pretraining, post‑training, multimodal and embodied intelligence—and reporting directly to Alibaba Cloud CTO and Tongyi Lab head Zhou Jingren (周靖人). This structure enabled efficient collaboration and rapid iteration, producing flagship models such as Qwen3.5.
Recently, Tongyi Lab planned to split Qwen into several horizontally divided teams—pretraining, post‑training, text, multimodal—still under Tongyi Lab but with Lin's management scope significantly reduced. Alibaba Cloud also introduced new leadership: former Google DeepMind senior researcher Hao Zhou (周浩), who participated in the Gemini series, was brought in to take over part of the post‑training work and possibly assume broader management responsibilities.
On the afternoon of March 3, Lin formally submitted his resignation. That evening, small‑scale internal messages circulated within the Qwen team, with several colleagues visibly distraught and some in tears. After Lin's public farewell in the early hours of March 4, Yu Bowen, Kaixin Li and other core members quickly followed with resignation announcements, creating a chain reaction.
Alibaba's senior management responded urgently. On the afternoon of March 4, Tongyi Lab held an all‑hands meeting attended by Alibaba Group CEO Wu Yongming (吴泳铭), Chief Talent Officer Jiang Fang (蒋芳) and Alibaba Cloud CTO Zhou Jingren. Wu acknowledged, "I should have known earlier." The meeting stressed that "Qwen is not shrinking but expanding," and announced the formation of a "Foundation Model Support Group" led by Wu and Zhou to coordinate group resources and increase investment.
Potential Consequences: Pressure on Qwen's Open‑Source Ecosystem and Alibaba's AI Strategy
Short‑term impact: slowed R&D rhythm and talent gaps
Lin and his team were the driving force behind Qwen's open‑source strategy. Their departures may temporarily slow Qwen's iteration cadence. Plans for a Singapore technical hub have reportedly been halted, and progress in multimodal and embodied intelligence directions faces uncertainty. The open‑source community has expressed concern about whether Qwen can maintain its global‑leading open momentum.
Mid‑term impact: challenges to organizational culture and morale
Some insiders interpret the restructuring as a clash between "KPI‑driven" and "technology‑driven" approaches—Alibaba Cloud's assessment of Qwen may shift toward commercial metrics like daily active users (DAU) rather than pure technical breakthroughs. This contrasts with Lin's insistence on "deep coupling across the large‑model full chain." Team morale has been dented, and future talent recruitment may become more difficult.
Long‑term consequences: a potential watershed for Alibaba's AI strategy
On the positive side, Alibaba has reiterated "continued commitment to open source and increased investment," and has formed a cross‑department task force, signaling the group's strategic emphasis on foundation models. Qwen remains a core component of the AI triangle linking Alibaba Cloud, Damo Academy and Pingtouge.
On the negative side, continued talent departures (previous losses include key Tongyi Lab figures such as Zhou Chang, Yan Zhijie and Bo Liefeng) could weaken Alibaba's competitiveness in open‑source large models. If the new teams fail to integrate quickly, Qwen may fall behind domestic rivals like ByteDance's Doubao and Zhipu's GLM, potentially eroding Alibaba's voice in the global AI ecosystem.
Industry takeaway
This episode again underscores the harsh reality of the large‑model era: talent is core competitive capital. Alibaba's internal adjustments reflect pressures across the industry—how to balance commercial KPIs with long‑term technicalism, and how to reconcile open source with commercialization.
Conclusion: Qwen's "soul departure" tests Alibaba's AI resolve
Justin Lin's resignation and the collective exit of core team members mark a major inflection point in Qwen's development. It is both the growing pains of organizational optimization and a necessary step as Alibaba's AI strategy moves from "rapid growth" to "refined governance." Whether Qwen can continue its open‑source legacy will depend on Alibaba's ability to quickly stabilize the team, rally new talent and maintain heavy investment in foundation models. Whatever the outcome, this event has become a significant footnote in China's 2026 AI story—an era of accelerated talent mobility in which every major player must confront the human variable.
▶️ Enjoyed this piece? Creating content is hard!
Follow our YouTube channel:
AI Rosetta
Your likes, subscriptions and shares are our greatest support! 🙏
More AI news at: https://www.airosetta.com