Responsible for discovery of insights in large-scale data sets, communication of insights to business partners, construction of statistical and machine learning models, and consulting on productionalization of these models and insights.Developed models for prediction of customer residential gateway failure for AT&T wireline products.• Developed time series forecasting methods for prediction of various technician metrics across AT&T technician types and geographies.• Developed method for prediction of nationwide market allocation across AT&T wireline products based on historical data.• Designed and built model for prediction of wireline jobs which are unlikely to be completed. Investigated approximately 50 features for potential inclusion in model, led calls with business unit partners on insights.• Designed algorithm for scheduling routine inspections of remote terminals to inspect low-risk RTs less frequently, enabling $20M in savings due to headcount reductions made based on it.• Built models for redeployment of technicians to small satellite work centers, leading to substantial reductions in mileage driven by technicians and enabling AT&T to monetize unnecessary real estate.• Analyzed impact of weather on AT&T network in order to identify network elements that are particularly affected by rain, to suggest them for rehab.• Designed and built model for predictive analytics on U-Verse network to proactively identify cables suffering from network issues. • Designed and built predictive model for identifying cell sites likely to be overpaid.• Filed two patents on improvements to U-Verse systems.• Coordinated hiring process and conducted interviews for data scientist candidates.• Led ideation sessions in which team of data scientists and engineers brainstormed new project ideas. Vetted new project ideas and presented them to business unit partners.

Quantitative analysis of secondary ticket markets and predictive analytics for pay-per-click advertising.Carried out quantitative analysis of secondary ticket markets for internal use and for consulting with major sports organizations.Improved algorithmic tools for trading strategy based on historical data, reducing need for human intervention and increasing revenue.Designed predictive modeling system for bidding on pay-per-click advertisements.Built tools for internal tracking of ticket sales.

Courses taught:Statistics 20, Introduction to Probability and Statistics, Fall 2010Statistics 21, Introduction to Probability and Statistics for Business, Spring 2011 and Spring 2012Statistics 131, Statistical Inference for the Social and Life Sciences, Spring 2012Statistics 134, Concepts of Probability, Spring 2011 and Fall 2011Statistics 135, Concepts of Statistics, Summer 2011Statistics 155, Game Theory, Fall 2010 and Fall 2011.Classes ranged in size from 40 to 100 students. Many of these classes had teaching assistants and undergraduate graders whose work I supervised.

Taught recitation sections and graded homework, quizzes, and exams for:Math 104: Calculus I, Fall 2006Math 114: Calculus II, Spring 2007 and Fall 2007Math 170: Ideas in Mathematics, Spring 2008Math 361: Advanced Calculus [real analysis], Fall 2009Graded homework for Math 501: Differential Geometry, Spring 2010.

Taught course in introductory calculus. Gave daily lectures, graded homework and exams.