LeiPhone has learned exclusively that Yao Xingcheng, who led pre-training for Tencent’s Hunyuan large language model, has joined North American start-up Thinking Machines Lab.
People familiar with the matter said Thinking Machines Lab offered Yao a package clearly above his Tencent income, which was already in the tens of millions of yuan per year. Earlier rumours had him joining Meta; the move instead landed at the AI company founded by former OpenAI chief technology officer Mira Murati.
Yao left Tencent in July. He led Hunyuan pre-training and was among the young AI research hires Tencent courted over the past year. He and Yao Shunyu both come from Tsinghua’s Yao Class and are academic brothers. LeiPhone had earlier reported domestic majors poaching pre-training talent from Gemini; now Tencent’s talent is flowing to an overseas lab.
The wider pattern matters more than one departure. For China’s big tech, top pre-training talent has become a scarce resource, and the leak shows the pull of US frontier labs with deeper pockets and a clearer path to frontier models. The competition for who trains the next model is now a competition for who can keep the trainer.
Editor’s note: This is an adapted translation of the original LeiPhone report. It has been trimmed and restructured for readability for an international business audience. The full original (in Chinese) is at https://www.leiphone.com/latest/index/id/4767.