Xu Jin graduated from the mixed class of Zhu Kezhen College of Zhejiang University, with a doctorate in signal and information processing. During the doctoral period, his main research interests were robot autonomous navigation, stereo vision, pattern recognition, path planning and machine learning.

One of the founders of Hangzhou Daily Technology Co., Ltd. (now Zhejiang Daily Interactive Network Technology Co., Ltd.), and later worked in Shanghai Research Institute of Huawei Technologies Co., Ltd., with rich experience in software and hardware development and team management. He is currently the technical director/fund manager of Magic Square Quantification, engaged in strategic research and responsible for the management of technical development team.

The Rubik's Cube is different from many domestic head quantification institutions in that its founding team has no working experience in large overseas quantitative hedge funds, and it is an outstanding representative who started from China and achieved outstanding results.

The development of magic square;

From 2008 to 20 14, the founding team of Rubik's Cube explored fully automated trading from scratch. The founding team of 20 15 is high-spirited, innovative and diligent, creates magic square quantification, makes quantitative investment by relying on mathematics and artificial intelligence, and is determined to become the top quantitative hedge fund in the world.

On June 20 16, 10, the first stock position generated by the deep learning algorithm model was actually traded online, and it was calculated by GPU. Prior to this, the algorithm mainly relied on linear model and traditional machine learning algorithm, and the model calculation mainly relied on CPU.

On 20 17, Rubik's cube continued to expand the AI algorithm research team and the AI software and hardware research and development team. By the end of 20 17, almost all the quantization strategies have been calculated by AI model.

On 20 18, the Rubik's Cube established AI as the main development direction of the company. The complex model computing requirements make single-machine training encounter the bottleneck of computing power, and the growing training requirements conflict with limited computing resources, so we seek a large-scale computing solution. In the same year, he won the Golden Bull Award for the first time.

Rubik's cube-Rubik's cube tutorial