I received my PhD from University of Michigan at 2012 and joined Google shortly after that. My research focused on automated diagnosis of software configuration problems that lead to crashes, undesired behavior, and also performance problems. My research interests, however, broadly include operating systems and distributed systems. I have been serving as program committee member in the following systems conferences:
Usenix ATC 2017
NSDI 2016 - LightPC
Co-chairing Diversity @ SOSP 2015
Recently, I've been shifting my focus to machine learning. Google has accumulated a lot of knowledge and expertise over the years in designing machine learning algorithms and applying them to a variety of domains. I am working with machine learning researchers to make Google-made machine learning available to people outside Google.
More specifically, I’m working with Natural Language Understanding researchers to identify the domains in which Google’s rich set of NLU models can help solve users’ problems. These domains include understanding sentiment expressed in text, understanding various topics and entities in a piece of text, and understanding low level language structures. Our users bring use cases from many different problem domains to us. Our goal is to design models and solutions to accommodate many use cases with high accuracy and flexibility.