Bio

Yang Zhilin is a Chinese computer scientist, born in 1992 in Shantou, Guangdong province. He is the co-founder and chief executive officer of Moonshot AI, the Beijing-headquartered foundation-model company he established in March 2023 with two Tsinghua University classmates, and a co-author of the Transformer-XL and XLNet papers from his doctoral period at Carnegie Mellon University. As of May 2026, he leads Moonshot AI following the April 2026 release of Kimi K2.6 and a reported $1 billion round at an $18 billion valuation.

Education

Yang entered Tsinghua University in 2011, initially admitted to the Thermal Energy Engineering program before transferring to the Department of Computer Science and Technology in his sophomore year. He completed the Bachelor of Science in computer science in 2015, advised during his undergraduate research by Tang Jie, the Tsinghua professor who later co-founded Z.AI / Zhipu AI.

He continued at Carnegie Mellon University's School of Computer Science for doctoral study, earning the PhD in 2019 under Ruslan Salakhutdinov and William W. Cohen. His dissertation, "Advances in Generative Feature Learning," was completed in four years against the CMU standard of six. The thesis covers his work on Transformer-XL, XLNet, and broader contributions to representation learning and language modeling.

Career

During the doctoral period Yang held research-internship positions at Google Brain with Quoc V. Le and at Facebook AI Research (later Meta AI) with Jason Weston. He co-founded Recurrent AI in 2016, a Beijing-based natural-language-processing startup focused on enterprise sales-conversation analysis. After completing the PhD he returned to China and worked with Huawei on early development of the PanGu language model series in 2020, then led work on the Wu Dao large language model line at the Beijing Academy of Artificial Intelligence in 2021.

He co-founded Moonshot AI in March 2023 in Beijing with Zhou Xinyu and Wu Yuxin, both former classmates from his Tsinghua undergraduate cohort. The three had been members of Splay, a rock band Yang joined as an undergraduate. The company name comes from Pink Floyd's album "The Dark Side of the Moon," and Moonshot launched on the album's fiftieth anniversary. The Kimi consumer chat assistant followed later in 2023, with long-context language modeling as the principal capability differentiator from contemporaneous Chinese competitors.

The Moonshot funding trajectory has been unusually rapid. The company raised at a $2.5 billion valuation in February 2024 and at $3 billion in September 2024, with prominent participation from Alibaba and HongShan among others. By March 2026 Moonshot was reported to have raised approximately $1 billion in a round valuing the business at $18 billion. The Kimi K2 release in July 2025, the natively multimodal Kimi K2.5 in January 2026, and Kimi K2.6 in April 2026 established Moonshot's open-weights credentials alongside DeepSeek and Alibaba Qwen as a leading Chinese frontier developer.

Notable contributions

Yang's published research record is concentrated in language modeling and representation learning, with a high-citation period at CMU during the late 2010s and a subsequent transition into industrial model leadership.

• Transformer-XL (January 2019). Co-author of the paper introducing recurrence into the Transformer architecture for long-context language modeling.
• XLNet (June 2019). Co-author of the autoregressive permutation-language-modeling method, which exceeded BERT on the GLUE benchmark at the time of publication.
• General Language Model (GLM) family. Contributor to the GLM line developed in collaboration with the Tsinghua group around Tang Jie.
• CodeGeeX code-generation model line, contributed during the Wu Dao period at BAAI.
• Moonshot AI founding (March 2023) and the Kimi product and Kimi K2 model line.
• Recognition. NVIDIA Fellowship; Siebel Scholar; Forbes Asia 30 Under 30; BAAI Young Scientist; named in TIME 100 AI 2024.

Affiliations

• Tsinghua University: Bachelor's student, 2011 to 2015.
• Carnegie Mellon University: Doctoral student, 2015 to 2019.
• Recurrent AI: Co-founder, 2016 to present.
• Google Brain: Research intern, doctoral period.
• Facebook AI Research: Research intern, doctoral period.
• Huawei Noah's Ark Lab: Researcher on the early PanGu series, 2020.
• Beijing Academy of Artificial Intelligence (BAAI): Lead on the Wu Dao large language model effort, 2021.
• Moonshot AI: Co-founder and Chief Executive Officer, March 2023 to present.

Position in the field

As of May 2026, Yang is the youngest chief executive among the principal Chinese frontier-tier AI lab leaders, alongside Liang Wenfeng of DeepSeek, Yan Junjie of MiniMax, and Tang Jie of Z.AI / Zhipu. The Moonshot strategy combines frontier-tier model research focused on long-context and agentic execution, open-weights distribution of flagship models under permissive licensing, and consumer distribution through the Kimi assistant in China.

The Kimi K2 release line through April 2026 has positioned Moonshot as one of the leading Chinese open-weights developers. Industry coverage has noted the K2.6 SWE-Bench Pro score of 58.6 as ahead of leading closed-weights coding competitors at the time of release. Whether the long-horizon agentic capability framing translates into commercial product wins against Anthropic Claude Code and OpenAI Codex is one of the most-watched questions for Moonshot in 2026.

Yang's public profile in Western press has grown alongside the Kimi K2 release cadence. The Wire China and South China Morning Post have profiled him in long-form coverage, and the Tsinghua-CMU academic background gives him a documented Western-academic record uncommon among his Chinese peer chief executives.

Sources

• Yang Zhilin. Wikipedia biographical entry covering education, the CMU doctoral period, the Moonshot AI founding, and recognition.
• Zhilin Yang personal homepage. Personal academic homepage with publication record, advisors, and research summary.
• Zhilin Yang Google Scholar. Google Scholar profile with the Transformer-XL and XLNet citation records.
• Zhilin Yang LinkedIn. LinkedIn profile.
• Who is Yang Zhilin, the Moonshot AI founder?. The Wire China long-form profile from September 2025.
• Meet Yang Zhilin, the Moonshot AI founder. Yahoo Finance / South China Morning Post profile.
• Transformer-XL: Attentive Language Models Beyond a Fixed-Length Context. The January 2019 paper introducing recurrence into Transformer language models.
• XLNet: Generalized Autoregressive Pretraining for Language Understanding. The June 2019 XLNet paper.