DR.
TILMAN DRERUP
ECONOMICS/MACHINE LEARNING
Welcome!
I am a Director of Machine Learning Engineering at Instacart in San Francisco, California. I lead a team of economists and MLEs working on a variety of topics, including ads, pricing, and growth.
This website contains a short overview of my past and present work and research. You can contact me through LinkedIn.
Updates:
April 2024: Levi Boxell, Robert Fletcher, and I have written another blog post to provide a closer look at our team's work. Check it out!
November 2023: Levi Boxell, Robert Fletcher, and I have written a blog post to describe the kind of work that Instacart's Economics team is working on and the tools and skills we deploy.
December 2022: George Gui and I have just released a new working paper on "Designing Promises with Reference-Dependent Customers: The Case of Online Grocery Delivery Time" - Check it out!
November 2022: Our paper on "Skewness Expectations and Portfolio Choice" with Matthias Wibral (Maastricht) and Christian Zimpelmann (Bonn) is forthcoming in the Journal of Experimental Economics.
April 2022: Update time! I have been promoted to Director of Machine Learning Engineering at Instacart!
January 2021: Our patent for a "Trustless Physical Cryptocurrency" has finally been granted.
July 2020: Exciting news! I'm joining Instacart as a Staff Economist in the San Francisco office.
October 2019: You can now check out what I have been working on prior to joining WeWork --- kong.cash is finally online.
August 2019: I have accepted a position with WeWork. I will be working as a Senior Economist at the office in Palo Alto, CA.
May 2019: My paper "Eliciting Subjective Expectations for Bivariate Outcomes" will be published in the Journal of Behavioral and Experimental Finance. Check it out!
01 PAST ROLES
2022 - present: Instacart
Director of Machine Learning Engineering (San Francisco, CA)
2020 - 2022: Instacart
Senior Engineering Manager, Machine Learning & Economics (San Francisco, CA)
2020 - 2020: Instacart
Staff Economist (San Francisco, CA)
2019 - 2020: WeWork
Senior Economist (Palo Alto, CA)
Leading the algorithmic pricing team, using a combination of machine learning and economics to recommend price adjustments for our office fleet.
2018 - 2019: Airtime Network / Kong
Economist / Co-Founder (San Mateo, CA)
Development of crypto-economic incentive systems, smart contracts, and mobile app for kong.cash.
2016 - 2017: Stanford University
Postdoctoral Researcher in Economics
Research under the supervision of Prof. Susan Athey. Focus on modelling user heterogeneity / dimension reduction techniques to improve predictions of standard multinomial choice models with a particular focus on estimation of price elasticities.
2013 - 2016: Bonn University
(Postdoctoral) Researcher in Economics
Research on empirical analysis of individual and collective decision making in financial markets. Teaching courses / lectures in mathematics, econometrics, corporate finance, and programming.
2003 - 2015: Bonn University / Bonn Graduate School of Economics
Diploma & PhD / Doctoral Research in Economics
PhD program at the Bonn Graduate School of Economics. Research focus on applied microeconometrics and the role of expectations in financial decision making.
2002 - 2006: Gameconomy GmbH
Founder / CEO
Platform for the sale of virtual goods in online role-playing games. Responsibilities included overall strategy, personnel, customer retention, advertising, and pricing. Received coverage on national TV and radio while I was CEO.
02 RESEARCH BACKGROUND
I have worked as a postdoctoral researcher in economics at Stanford University under the supervision of Prof. Susan Athey and as a postdoctoral researcher at the Bonn Institute for Applied Microeconomics.
My research interests include econometrics, machine learning, applied microeconomics, and behavioral finance. Recently, I have been working on user heterogeneity / dimension reduction techniques in multinomial choice models with a particular focus on the estimation of price elasticities. I'm also interested in the role of beliefs and non-standard choice behavior in financial decision making. Below is a list of selected current and past projects.
SELECTED PAPERS
[2022] Designing Promises with Reference-Dependent Customers: The Case of Online Grocery Delivery Time
Joint Work with George Gui (Stanford)[2023] Skewness Expectations and Portfolio Choice
Joint work with Matthias Wibral (Maastricht) and Christian Zimpelmann (Bonn)
Journal of Experimental Economics, Volume 26, March 2023, Pages 107-144.[2019] Eliciting Subjective Expectations for Bivariate Outcomes
Journal of Behavioral and Experimental Finance, Volume 23, September 2019, Pages 29-45.[2017] The Precision of Subjective Data and the Explanatory Power of Economic Models
Joint work with Benjamin Enke (Harvard) and Hans-Martin von Gaudecker (Bonn)
Journal of Econometrics, Volume 200, Issue 2, October 2017, Pages 378-389.[2015] Diurnal Rhythms in Investor Sentiment
Journal of Behavioral and Experimental Finance, Volume 7, September 2015, Pages 71-81.[2014] Recommendation Revisions and Differences in the Interpretation of Earnings
03 CODING
I have more than 10 years of coding experience in both research and corporate settings. I work in a variety of coding environments, including Python, R, Stata, JavaScript (NodeJs, React Native), Solidity, as well as basic bash and SQL.
I maintain the following open source repositories:
ELLIPTIC CURVE SOLIDITY
https://github.com/tdrerup/elliptic-curve-solidity
A Javascript / Solidity implementation of elliptic curve math. Parameterised to enable ECDSA signature verification for curve secp256r1 / prime256v1 / p256 on the Ethereum blockchain.
BINARY CHOICE DOUBLE INDEX
https://github.com/tdrerup/binary_choice_double_index
A Python implementation of the semiparametric double-index estimator of Klein and Vella (2009).
I also wrote the following mobile app (iOS / Android) during my work with Airtime Network:
KONG.CASH MOBILE APP
https://apps.apple.com/app/id1479791375
The app verifies physical crypto notes using a combination of smart contracts, elliptic curve math, and NFC.