When I was 12, I somehow became obsessed with the Enter key on keyboards; I was wondering what it does and how to use it. My parents bought me an “educational console” (a Famicom clone with a keyboard); on the console, I typed (copied from a manual) my first program, which was a Mario game written in BASIC. That was the moment everything started.

In middle school, my parents bought me a PC for 3,900 CNY (~550 USD); we upgraded the RAM to 64MB for another 500 CNY (~70 USD). I was playing with HTML, CSS, ASP, etc. before college. I sold a few websites to companies/hospitals. I also participated in the Olympiad in Informatics in high school.

In college and grad school in China, I focused on embedded systems because of a random blog post online saying that’s the future. I was playing with assembly, low-level C, communication protocols (e.g., USB), and so on. In college, our team of three students built a system called “Interactive Window”, which allowed users to interact with museum displays using hand gestures; we won the first prize (top 13 of 160 teams) in the National Undergraduate Electronic Design Contest. Later, in grad school in 2009, I built the LiDAR solution for our school’s self-driving car; the team won the first place in the first National Self-Driving Vehicle Contest.

Then I came to the US for a PhD, focusing on data infrastructures. I published at conferences like HPDC, FAST, and EuroSys. Traveling for conferences was like the reward for the work :) In the early years in Wisconsin, I was obsessed with statistics. I got statistics as my minor and wrote a paper using statistical methodologies to analyze Linux kernel file system properties. I liked running experiments and visualizing data using ggplot2 in R.

Then I joined Google, working on large-scale data infrastructures (F1 Query team). The first a few years were about sort-based query optimization. In the past a few years, I have been leading a team to help Google manage data better on Google’s distributed file system for better usability, governance, and performance.