Welcome! I am an economist at Amazon on the Topline Forecasting team. We develop methods to forecast Amazon's demand. My interests lie at the intersection of econometrics and machine learning. I have a particular interest in explainable artificial intelligence. We live in a world where vast new datasets are becoming available every day. This lets us rigorously identify statistics that summarize the key information in the data for decision makers in real time. Against that background, I develop parsimonious nonparametric forecasting methods at the intersection of traditional time series econometrics and machine learning.
In Jumps, Realized Densities, and News Premia, I build an interpretable nonparametric framework relating high-frequency and daily returns. This framework provides a novel sufficient statistic for time-varying news risk — jump volatility — which I estimate using high-frequency data. In Feasible Multivariate Density Estimation using Random Compression, I develop Bayesian methods to feasibly estimate the joint density of 5 to 10 variables and provide new theory to explain why this is actually doable. In Identification Robust Inference for Risk Prices in Structural Stochastic Volatility Models, I provide identification-robust inference for the prices of market and volatility risk in the presence of stochastic volatility and leverage effects.