I am interested in building an integrated machine learning system that is general enough to represent sufficient knowledge, robust enough to tolerate noises, and fast enough to scale up and scale out. Specialties: Statistics; Machine Learning; Optimization algorithms; Artificial Intelligence.

Senior Software Engineer @ Online Relevance Infrastructure From October 2014 to Present (1 year 2 months) Mountain View, CAApplied Researcher @ Facial expression understanding for xbox one. Personalized machine learning system: A distributed machine learning system for automatic personalization through Alternating Direction Method of Multipliers (ADMMs) Machine learning APIs: compile-time optimization as machine learning service. From October 2012 to September 2014 (2 years) RedmondResearch Associate @ Research in machine learning on building an integrated interactive learning system based on large scale probabilistic graphical models. From August 2011 to October 2012 (1 year 3 months) Research Assistant @ Research Large scale parallel global optimization 2010-2011 Margin based object networks 2010-2011 Interactive learning for semantic video analysis 2009-2010 Large margin sigmoid belief networks 2008-2009 Large margin Boltzmann machines 2008-2009. Complementary kernel density estimation 2006-2008. Learning the Lie groups of visual invariance 2005-2006. From January 2005 to August 2011 (6 years 8 months) Visiting Scholar @ With a couple of scientists, we are developing a large scale global optimization system for tuning expensive systems. The system is parallelized to scale up with high dimensional problems. The optimization algorithms are designed to handle complicated non-convex functions. From July 2010 to October 2010 (4 months) Visiting Researcher @ As a visiting researcher, I joined a team to work on the high-level feature extraction task for broadcasting videos. I proposed and implemented the Large Margin Sigmoid Belief Networks to model the semantics of the high-level features, and applied the interactive learning approach to train the model. Basically the model can communicate with human operator through an efficient user interface. From July 2009 to September 2009 (3 months) Teaching Assistant @ Teaching Java programming language and data structure. From October 2003 to January 2005 (1 year 4 months)

Ph.D, Computer Science and Engineering @ University of Washington From 2006 to 2011 M.A, Computer Science and Engineering @ University of Washington From 2003 to 2006 B.Eng, Electronic Engineering @ Tsinghua University From 1995 to 2000 Xu Miao is skilled in: C++, Java, Python, Machine Learning, Theory, C, Algorithms, Linux, Statistics, Scalability, Artificial Intelligence, Programming, Computer Vision, Distributed Systems, Computer Science, Software Engineering