Meet Yang Zhilin, the mind behind China's Kimi K3 model
What's the story
The launch of Kimi K3, a powerful new AI model from China, has put the spotlight on Yang Zhilin. The Moonshot AI co-founder is being hailed as one of the brightest minds in China's AI ecosystem after his start-up unveiled this groundbreaking model. Here's everything you need to know about him and his journey in the world of artificial intelligence (AI).
Education
From Tsinghua to Carnegie Mellon
Zhilin studied at Tsinghua University before moving to Carnegie Mellon University in the US, one of the world's top AI research institutes.
He completed his PhD in just four years under renowned AI researcher Russ Salakhutdinov.
His academic work earned him a reputation as a rising star in machine learning, with Salakhutdinov recently describing him as "absolutely brilliant" while congratulating him on Kimi K3's success.
Career path
His career in AI and pivotal decisions
After his academic stint, Zhilin worked at tech giants like Google Brain and Meta. However, instead of pursuing a long-term career in Silicon Valley, he returned to China.
There, he contributed to several major AI initiatives, including projects related to Huawei and the Beijing Academy of Artificial Intelligence.
The story of Apple wanting to recruit Zhilin, which emerged after Kimi K3's release, shows his desire for independence over corporate success.
AI perspective
His philosophy on AI development
Zhilin believes that scaling AI systems can be more effective than constantly searching for new algorithms.
He has said, "If you can solve it with scale, don't solve it with a new algorithm."
This philosophy is reflected in Kimi K3, which packs an enormous number of parameters and a massive context window to process huge volumes of information and tackle complex reasoning tasks.
Start-up journey
About Moonshot AI and its rise
Founded in 2023, Moonshot AI has quickly become a major player in the global AI race. The start-up is backed by major investors and reportedly valued in the tens of billions of dollars.
Unlike many American rivals, it embraces open-weight AI models that developers can download, modify and run themselves.
This strategy has won support among researchers and open-source advocates while helping Moonshot emerge as a serious challenger to established AI giants.