Work Experience

Current	Staff Research Scientist Lyft. Member of Rideshare Labs: Experimentation and Causal Inference where I focus on developing novel, scalable, causal inference methods to infer long-term outcomes from short term observational signals.
Current	Machine Learning Researhch Engineer Conde Nast. Implemented a trending articles section on Vanity Fair which incorporates of- fline and online reinforcement learning methods. Using transfer learning via optimal transportation to improve performance of an object detection model for Vogue and Vanity Fair.
2016-2018	Data Scientist. Conde Nast. Developed and productionalized acquisition and personalization models for The New Yorker. Constructed A/B tests to measure performance along with a reporting server updated daily to ensure stability.
Current	Adjunct Professor of Data Science. Columbia University. Course: APME4990: Introduction to Data Science in Industry Github: https://github.com/Columbia-Intro-Data-Science/APMAE4990-
2014-2016	Data Scientist. The New York Times. Developed and productionalized a churn model API which was used in the Sugar Care call center to inform agents how to treat customers. Developed a customized time series model for delivery of papers to Starbucks, which was implemented for all stores across the United States.
2013-2014	Herchel Smith Fellow and Lecturer of Pure Mathematics. Cambridge University.

Education

2013	Courant Institute, NYU & Universite Paris VI Pierre et Marie Curie Paris. Ph.D. (NYU), Mathematics. Thesis: "Energy driven pattern formation in a non-local Ginzburg Landau energy.". Doctorat (Paris VI), Applied Math. Advisor: Sylvia Serfaty.
2008	University of Toronto. M.S., Mathematics. Thesis: "Weak Lagrangian solutions to a one dimensional version of the semi-geostropic equations with moisture.". Advisor: Robert McCann.
2007	University of Toronto. B.Sc, Mathematics and Physics Specialist.(with high distiction) Thesis: "Chaotic response of the 2d semi-geostropic equations to mild periodic perturbations.". Advisor: Robert McCann.

Skills

Technical	SQL Hive Hadoop Google Cloud Platform AWS GCP Druid Python Scikit-learn Bash Linux/Unix
Machine Learning	General Linear Models Ensemble Methods: Random Forests, Decision Trees, Gradient Boosting Recommendation Engines: Graph Diffusion, Matrix Factorization Clustering Methods. A/B tests and Causal Inference Bayesian Inference. Offline and Online Reinforcement Learning Monte Carlo Sampling Methods Deep Learning/Transfer Learning Deep Learning and Transfer Learning