The organizational restructuring in the core technology areas of artificial intelligence at Baidu is still deepening. It is reported that after Sun Tianxiang officially took office as the head of the Baidu Foundation Model Research Department (BMU) in July, a new round of systematic restructuring and upgrading of the organization and technical teams around the ERNIE large model has been carried out.
In this adjustment, Baidu welcomed two key technical experts. On one hand, Wei Haoran, a former researcher from DeepSeek, has officially transferred his team to join the Baidu Foundation Model Research Department, and will serve as the head of the ERNIE multimodal algorithm. Wei Haoran has shown outstanding performance in his previous research experience, accumulating rich technical achievements in large language models and other cutting-edge fields such as optical character recognition (OCR). On the other hand, Baidu successfully introduced Wu Qin, a senior large model researcher who has participated in the development and iteration of multiple generations of large models, from a well-known Frontier Lab in North America, aiming to comprehensively strengthen and consolidate the moat of the ERNIE large model in pre-training capabilities.
Looking back at the high-level personnel changes since July this year, after Sun Tianxiang joined Baidu as the head of the Foundation Model Research Department, he also joined the Baidu Model Committee. Moreover, in late July, as a 90s-generation technical expert, he officially entered the newly restructured Baidu Technical Strategy Committee, fully leading and promoting Baidu's strategic layout and technical breakthroughs in the direction of foundational large models.
