In Profile: The Alibaba Engineer Behind Qwen Who Walked Away

1. Lin Junyang, a prominent technical lead who played a significant role in developing Alibaba Group’s Qwen open-source AI model series, unexpectedly departed from the company in early March 2024. His exit, along with several other Qwen team members, sent shockwaves through China’s AI industry and exposed concerns about talent retention at one of the country’s biggest technology firms. Lin’s departure triggered widespread speculation and anxiety within the sector, given his status as one of Alibaba’s youngest high-level technical experts. [para. 1][para. 2][para. 4]

2. Lin’s resignation was announced through a brief post on X, and his subsequent emotional statements revealed the collective impact his work had on the community. The sudden exit was seen as abrupt within Alibaba Cloud, with employees acknowledging a likely short-term impact on the Qwen project. However, the company quickly reaffirmed its commitment to AI development and its open-source strategy, suggesting that long-term effects would be mitigated if resource investment continued. [para. 2][para. 3][para. 5][para. 6]

3. Lin’s background is notable for its interdisciplinarity. He studied English as an undergraduate and later specialized in foreign linguistics and applied linguistics at Peking University, before transitioning towards the intersection of language and computer science during graduate school. This unique combination enabled him to excel in natural-language processing and contribute significantly as Alibaba moved towards large foundational AI models. His experience bridged linguistics and engineering, sharpening his ability to work on complex large language models. [para. 7][para. 8][para. 9][para. 10]

4. Lin joined Alibaba’s Damo Academy in 2019, initially contributing to natural-language processing for applications like search and recommendation. As Alibaba shifted focus to foundational models, Lin’s technical groundwork became fundamental for the multimodal and large-scale model architecture underlying Qwen. In 2022, as generative AI became a central tech battleground, Alibaba consolidated its AI efforts under the Tongyi Lab, appointing Lin as head of the Qwen project. He spearheaded the launch of the Tongyi Qianwen model in April 2023, helping Qwen become one of the world’s most downloaded open-source model series, with over a billion downloads and more than 200,000 derivative models. [para. 10][para. 11][para. 12][para. 13][para. 14][para. 15]

5. Lin advocated for the practical application of large models and a vision driven by AI’s real-world utility. He recognized the natural divergence in AI companies’ strategic focuses—catering to enterprise, consumers, or specialized domains. In 2025, Lin’s achievements earned him a P10 promotion at age 32, one of Alibaba's top technical ranks. However, insiders noted that, while Lin was a prominent technical leader, the open-source strategy was defined by Zhou Jingren, Alibaba Cloud’s CTO, rather than Lin alone. [para. 16][para. 17][para. 18][para. 19][para. 20]

6. The timing of Lin’s departure was unexpected, coming days after the launch of the Qwen3.5 small-model series, which received positive attention from figures like Elon Musk. The split reportedly arose from disagreements over strategy and lab structure, such as Lin favoring integration of pretraining and post-training, versus market trends toward their separation. Lin also wanted greater independence for the Qwen team’s AI infrastructure. These disputes reflected broader shifts at Alibaba Cloud towards applying AI to vertical industries and prioritizing productization, sometimes at odds with foundational AI research. [para. 21][para. 22][para. 23][para. 24][para. 25][para. 26]

7. Lin’s exit is part of a larger talent exodus from Tongyi Lab over the past two years, which included high-profile departures to rivals such as ByteDance and Meta. Multiple Qwen engineers left in parallel with Lin, reflecting the high competition for AI talent and the complex internal dynamics as Alibaba scales its AI efforts. New leadership, such as Zhou Hao from DeepMind, has stepped in to fill key roles. [para. 27][para. 28][para. 29][para. 30][para. 31]

8. In recent industry commentary, Lin expressed views on the shift toward agent-style AI applications and autonomous learning, suggesting that future systems might act proactively without human prompts. He gauged China’s chances of producing a world-leading AI company in the next three to five years at only about 20%, due to investment and research gaps, and emphasized the challenge of retaining top talent as recruitment competition intensifies. The departure underscores the ongoing challenge for Alibaba and other Chinese tech giants to sustain growth and innovation amid fierce global competition for AI experts. [para. 32][para. 33][para. 34][para. 35][para. 36]