Experience

IRL - Social Messenger

• United States
• Consumer Services
• 1 - 100 Employee

• Director of AI

  • Nov 2021 - Present



Amazon

• United States
• Software Development
• 700 & Above Employee

• Sr Applied Scientist / Tech Lead

  • Dec 2019 - Nov 2021



A9.com

• United States
• Technology, Information and Internet
• 100 - 200 Employee

• Applied Scientist / Tech Lead

  • Sep 2015 - Nov 2019

  * Leading a team to build pipelines to monitor ads response prediction model metrics * Led a team to model uncertainties of estimated click through rate of Amazon ads and evaluated the model in prdocution for reinforcement learning * Analyzed interactions between Amazon organic search and advertisement * Designed and conducted A/B experiments to evaluate models for Amazon ads * Leading a team to build pipelines to monitor ads response prediction model metrics * Led a team to model uncertainties of estimated click through rate of Amazon ads and evaluated the model in prdocution for reinforcement learning * Analyzed interactions between Amazon organic search and advertisement * Designed and conducted A/B experiments to evaluate models for Amazon ads



23andMe

• United States
• Biotechnology Research
• 400 - 500 Employee

• Scientist

  • Jul 2013 - Sep 2015

  * Analyzed 23andMe's large genetic database: millions of customers and genetic variables and thousands of self reported health variables for each customer * Provided statistical guidance for 23andMe's commerical and academic research collaboration studies such as the genetics of chronic pain, mosquito bite size, gestation length and etc. * Authored the discovery paper of morning person genetics at Nature Communications based on the analysis of 23andMe cohort (the discovery has been reported by many media sites)



University of Michigan

• United States
• Higher Education
• 700 & Above Employee

• Research Assistant

  • Sep 2008 - Aug 2012

  * Led the analysis of the genetic associations of lipids in a large scale NIH project: 6,000+ samples, each with >30 millions genetic variables * Developed a hidden Markov model based method and implemented it in C++ to infer the fine scale ancestries of African American * Developed a method and implemented it as a software to calcualte sample size for longitudinal data with binray outcomes and demonstrated sample size savings via a novel approach * Led the analysis of the genetic associations of lipids in a large scale NIH project: 6,000+ samples, each with >30 millions genetic variables * Developed a hidden Markov model based method and implemented it in C++ to infer the fine scale ancestries of African American * Developed a method and implemented it as a software to calcualte sample size for longitudinal data with binray outcomes and demonstrated sample size savings via a novel approach