The field of AI large models has once again received significant news. Zhang Xiangyu, a distinguished scientist in the AI domain and one of the four authors of the ResNet study, has joined the domestic large model unicorn enterprise, Jieyue Xingchen, drawing widespread industry attention.
Zhang Xiangyu's academic achievements are remarkable. In 2008, he was admitted to Xi'an Jiaotong University, later switching to the software engineering major. In 2011, he received a special prize nomination in the Mathematical Contest in Modeling for American college students, setting the best record in Xi'an Jiaotong University's history. In 2012, Zhang Xiangyu earned his bachelor's degree in software engineering and was selected for the joint doctoral training program between Xi'an Jiaotong University and Microsoft Research Asia, under the tutelage of Sun Jian and He Kaiming.
During his time at Microsoft Research Asia, Zhang Xiangyu, along with He Kaiming, Sun Jian, and Ren Shaoqing, completed the landmark ResNet research, where he was responsible for the underlying framework and coding. The core idea of ResNet—residual connections—has become an indispensable component of modern deep learning models. Since its publication in 2016, the ResNet paper has been cited over 230,000 times, making it one of the most popular deep learning papers of the century.
Zhang Xiangyu's addition marks Jieyue Xingchen's further commitment to the AI large model field. His research interests include the design of high-performance convolutional networks, AutoML, automated neural network architecture search, and the pruning and acceleration of deep models. In 2017 and 2018, Zhang Xiangyu completed another notable work—the mobile-efficient convolutional neural network ShuffleNet, which has had a profound impact on millisecond-level facial unlocking technology for smartphones.
In addition to Zhang Xiangyu, Jieyue Xingchen has also attracted other top talents in the AI field. Yu Gang from Tencent and Duan Nan from Microsoft Research Asia have also joined Jieyue Xingchen, injecting new vitality into the company's R&D capabilities.
Yu Gang served as the research director at Tencent and previously led the Detection team at Megvii. He has achieved significant accomplishments in research and algorithm implementation in detection, segmentation, tracking, skeleton, and action behavior. Duan Nan, a senior principal researcher at Microsoft Research Asia, focuses on natural language processing, multimodal intelligence, code intelligence, and machine reasoning.
Jieyue Xingchen's appeal lies not only in its status in the AI large model field but also in its talent strategy and technological innovation. The company's founder and CEO, Jiang Daxin, a former global vice president at Microsoft, has attracted many top talents in the AI field with his leadership and vision. Jieyue Xingchen's self-developed models and products, such as the Step series of general large models and the Yuewen & Maopao Duck, have also gained widespread recognition in the market.
As AI technology continues to evolve, the competition in the large model field is becoming increasingly fierce. With its strengths in talent and technological innovation, Jieyue Xingchen is poised to gain a favorable position in this competition. We look forward to Zhang Xiangyu and other AI experts bringing new breakthroughs and developments to Jieyue Xingchen.