Jingyue Wu

Jingyue Wu (吴景岳)


1600 Amphitheater Parkway
Mountain View, CA 94043

Email: <last><first> AT gmail DOT com

About me

I am building Waymo's machine learning infrastructure. Before transferred to Waymo, I wrote compilers at Google for machine learning.

  • gpucc: an open-source, fully functional, high performance CUDA compiler. It has powered numerous GPU applications at Google and is integrated into LLVM. I presented this work on behalf of my team in LLVM Dev 2015, CGO 2016, and GTC 2016. The slides can be found here.
  • XLA: a domain-specific compiler for linear algebra. It presently lives under TensorFlow and accelerates TensorFlow computations.

I completed my PhD in Computer Science at Columbia University. I worked with Professor Junfeng Yang on several projects related to software reliability and programming languages. During my years at Columbia, I also interned at Microsoft Research Silicon Valley and Facebook working on DryadLINQ and HHVM respectively.

I used to be an active programming competitor. I won the fifth place in China National Olympiad in Informatics (Chinese equivalent of USACO) back in high school. I attended ACM/ICPC, Topcoder, and Google Code Jam in my college and graduate school years. I also served as the coach for Columbia's programming contest team.

Selected publications

Papers I led

Papers I helped with

  • 09/2008 - 10/2014: PhD, Department of Computer Science, Columbia University, New York, USA
  • 09/2006 - 08/2008: Andrew Yao's Special Class (only 24 of all undergraduate students at Tsinghua University selected)
  • 09/2004 - 08/2008: BE with honors, Department of Computer Science and Technology, Tsinghua University, Beijing, China