Oct 2017 – Nov 2022 · 5 yrs 1 mo · Menlo Park, California

• staff scientist/tech lead manager Meta's private ads experimentation platform. This platform enables accurate and high-throughput A/B tests on ads delivery (e.g. a new ML model to rank ads) in a way that's compliant with various privacy restrictions. It is expected to facilitate ~30% of company-wide improvements in ads personalization in 2023 and eventually get to ~100%.
• Efficient experimentation to run fewer and better experiments via offline evaluation (e.g. meet 1.5% of Ads ML annual goals in 2019) and sample-efficient optimization (e.g. increase Instagram revenue in 2020 by ~1%)
• I am typically the first data scientist on high risk-return problems that require an interdisciplinary skillset and an ability to put on a research scientist, data scientist, and product manager hats.

Jun 2016 – Sep 2016 · 3 mos · Menlo Park, California

• Co-developed simple, broadly applicable and scalable statistical techniques to shorten A/B tests at Facebook.

Aug 2016 – Aug 2017 · 1 yr · Mountain View

• Dynamic discounts for restaurants in Palo Alto, Mountain View and San Francisco

Jan 2016 – Mar 2016 · 2 mos

• First year Econ PhD econometrics course, Outstanding Teaching Award based on student evaluations

Jul 2013 – Sep 2017 · 4 yrs 2 mos

• Research assistant to Frank Wolak. Worked with large-scale distributed data tables and used various numerical libraries and computational tools to conduct statistical estimation, simulations and empirical market design

Aug 2012 – Sep 2017 · 5 yrs 1 mo

• Dissertation on sequential hypothesis testing & implications of ride-sharing on cabs in the US

Sep 2010 – Apr 2012 · 1 yr 7 mos · Princeton, NJ

• Game theoretic & empirical work on ethic/religious polarization in the Middle East

Jun 2009 – Jul 2009 · 1 mo

• Preparatory work on various mergers and acquisitions