I love taking the latest and greatest technological innovation and creating products that provide tangible value for the user. Leveraging my deep roots in science and research, I create revolutionary products that solve real problems in people's daily life.I am currently a post-doctoral research fellow at the University of Michigan focusing on building the next generation cloud and mobile computing system for AI applications and workload. I am also building Jaseci, an open-source computational model, technolodgy stack and programming methodolodgy for bleeding edge AI. The mission of Jaseci is to enable any developers to build next-generation bleeding edge AI products at scale rapidly with zero AI, backend or DevOps experience.Previously, I managed the Clinc AI Platform, the world’s most sophisticated conversational AI platform, where he is productizing the latest research innovation in Natural Language Processing (NLP) to enable anyone to build the most natural and “human-in-the-room” conversational AI experience, effectively democratizing AI. The Clinc AI Platform is the flagship product of the company and is now serving more than 7 million users worldwide.While building the Clinc AI Platform, I continued to invent new ways to effectively leverage AI technology for real-world problems. My work in training data collection for intelligent applications takes a deep look at leveraging the public crowd to collect large corpus of training data quickly for building state-of-the-art NLP models, examining the bottleneck in the current common practices and designed new best practices, which have been employed at Clinc to this day and have been one of the core driving principles for Clinc’s various Virtual Assistants products.I earned my Ph.D. from the University of Michigan, working under the supervision of Professor Lingjia Tang and Jason Mars. My research focused on hardware and software system design for emerging intelligent applications, such as Apple Siri and Amazon Alexa. Through identifying the various inefficiencies that exist across the system layer, in both the hardware system design and the machine learning model training process, I designed and implemented optimization techniques to improve software system performance and machine learning performance, in the interests of providing a more natural and more responsive end-user experience as this class of applications grows in popularity and maturity.