I lead Amazon Search's data science group. We create the science and infrastructure behind metrics and experimentation that allows Amazon to build a world class search and discovery experience. We research how search results, presentation, and interface impact the way users shop, and the way sellers find buyers. That is, we are valuing the core premise of the internet's ability to reduce informational frictions (for commerce, anyway). To this work, we bring expertise in constrained optimization, causal inference, and marketplace equilibrium effects. To look around corners on how customer behavior is evolving in e-commerce, we run experiments and otherwise serve as the 'front line recon' on potential new ways for making shopping work for customers.... all at petabyte scale.

A9 was created to bring world class search to Amazon's millions of customers. I came to A9 to bring an economic lens on shopping to Amazon Search.

Led a small band of PhD-trained economists who use economic theory and econometrics to help eBay manage a global multi-billion-dollar marketplace. We engage with teams ranging from product engineering to HR to finance, wherever there may be a problem that can benefit from thinking like an economist. Market experiments and economic models of platforms in a big data environment deliver business objectives and better market outcomes. Just a few examples of work to date:- Created new product features inspired by behavioral economics that raised conversion rates for eBay's bargaining platform (Best Offer)- Demand estimation algos for optimizing return on targeted price subsidies, to automate manual deals curation by merchants- Estimated the benefits and costs of shipping forecasts for conversion gains balanced against the derived consequences of late delivery - Created optimal taxation models, and model calibration through elasticity estimates, for platform pricing- Implemented machine learning heterogenous treatment effects modeling for targeted advertising engine- Measurement of long term consequences of platform choices using econometric models and AB tests to understand outcomes over longer horizonsWe also foster academic understanding of the economics of e-commerce through collaboration with researchers in academia and publication of original research. Many of the eBay papers in top economics journals are results of collaboration with our group.

As an economist and researcher at eBay, I engaged with a wide range of teams to use economics and econometrics to meet business objectives and generate economic research.Some examples of my impact:- Designed innovative experiments to measure causal (incremental) returns to marketing in areas of paid search, display, affiliates, deals promotions, and even TV. Results were used to retarget more profitable ad demographics and raise profitability.- Executed large scale cross-channel test to calibrate a Media Mix Model for optimal balancing of marketing spend- Built a multi-touch attribution system using large scale logistic regression models which can generate predicted values for tracking real-time partial contributions of internet marketing channels.- Created a machine learning econometric estimator (a propensity score estimator with a random forest first stage) for estimating the value of a rewards program- Leveraged bargaining theory to guide offers on eBay's Best Offer feature which proved to increase conversion rates- Used NLP and machine learning to mine text in negotiation communications to improve conversion success odds (see working paper on my website)- Modeled AB testing on a platform (marketplace) and created heuristics for understanding the biases caused by market interactions (see publications below)- Estimated costs of searching for items in e-commerce (see publications below)This work has informed eBay market design to achieve business objectives as well as led to publications of original research.

Provided economic damages calculations for expert testimony in the areas of intellectual property, anti-trust, consumer class action, employee class action, and fraud. Specialized skills in data analysis and statistics.