It is reported that Luxudong Lin, the head of XAI's multimodal team, has joined Tencent as the head of Huan Yuan's multimodal algorithm

According to insiders, Luxudong Lin, the head of multimodal at Musk's xAI, has recently joined Tencent as the head of the multimodal content generation algorithm for Huan Yuan.

According to the information, Luxudong Lin graduated from Tsinghua University in 2018, then pursued a Ph.D. at Columbia University, with research areas including embedding learning, video analysis, and generative models. During his doctoral studies, he interned at Tencent, Facebook AI, and Google, and as the first author, he participated in the development of the Vx2Text framework, which enhanced the ability of subtitle generation through multimodal fusion.

After obtaining his Ph.D. in 2024, Luxudong Lin joined Google DeepMind and participated in the multimodal pre-training and post-training of Gemini. In November 2025, he joined xAI as the head of multimodal understanding, responsible for multimodal content generation, understanding, and training of generative models.

This year, Tencent Huan Yuan has been continuously adjusting its team. On July 7, Tian Yonglong, a researcher from OpenAI, joined Tencent as the head of multimodal understanding; on July 24, the departments of Huan Yuan's large language model and multimodal model were merged, led by Yao Shunyu. With the addition of talents such as Tian Yonglong, Pang Tianyu, and Luxudong Lin, the core team of Huan Yuan is basically formed, and the multimodal capability may become the key direction of future development.