Angie.H Moon develops a workflow that balances and tests the needs of stakeholders, the structural knowledge of experts, and machine learning methods. One goal is to automate this cycle using a modular architecture so that its product, explainable policy, can be broadly distributed. She is interested in end-to-end risk control tools for retailers which motivated her previous contributions to startup (NextOpt CEO), education (textbook Bayesian Data Analysis translator), and open-source (Stan, Arviz, SBC). She is a Ph.D. at MIT, exploring operations and entrepreneurship advised by Charles Fine. Previously, she was a pre-doc working on estimation in a dynamic model with Hazhir Rahmandad at MIT Sloan’s System Dynamics. She finished her master’s at Columbia University and undergraduate at Seoul National University, both industrial engineering majors.
Resume is here (updated 01.25.2023) but stories untold is more fun:
Her research on Bayes taught her to live by the spirit of gradual updates with simulation-based reliability checks. After identifying Bayesian offers the most practical framework for demand modeling with three years of experiments, she devoted five years to Bayesian and dynamic model methodologies (e.g. this paper advised by Andrew Gelman, TAing seminar Bringing data into dynamic models). With this methodological background, she is trying to find a research area that can have a real-world impact.
Detailed research can be found here.