At UC Berkeley, he was engaged in the joint research of DARPA-supported global speech recognition: EARS PROJECT, and took the lead in using multi-layer artificial neural network to promote the accuracy of speech recognition and made a breakthrough: in 2004, he ranked first with IBM in the DARPA evaluation, and was recognized by American Extraordinary Ability talents. Qingrui was founded in 2011. It combines AI technology with foreign language teaching and develops a unique third generation AI technology: brain-like cognitive computing model to better serve foreign language teaching. At the same time, leading technologies such as image recognition, speech synthesis, composition correction, natural language understanding and man-machine dialogue are introduced to build an intelligent listening and Speaking Coach who can listen and speak.