As Principal Data Scientist at Lumo, I am leading efforts on the development of novel machine learning and analytics methods for predicting flight delays due to inclement weather and other disruptive events. Predictions are made all the way from months in advance down to an hour before scheduled departure. Lumo is used by individual travelers, corporate travel planners and the airline industry.

At Aetna I led a team of 3 Data Scientists in the development of predictive models of health insurance claims, given various demographic, geographic and insurance provider features. The work involved investigating dimensionality reduction algorithms (PCA & ICA) to the original large data sets and then building predictive models for the insurance claims using machine learning algorithms.

Sonde Health is a digital medicine company developing a voice-based technology platform with the potential to transform the way we monitor and diagnose mental and physical health. Our products are based on proprietary, best-in-class technology developed by experts from leading academic institutions. Analysis of subtle but characteristic changes in your voice can accurately reveal important aspects of your health. Sonde's platform is designed with the privacy of the user in mind, and will work on devices most people already own, preventing you from having to wear something new or add anything to your busy routine.Our goal is to build new ways to follow mental and physical health to help you and your care team understand your health more completely so you know when and how to get the right help. Together, we have the potential to advance our understanding of what it means to be and feel well.

Working on applications of machine learning, network modeling & analysis, and mathematical & statistical modeling to problems in aviation and transportation systems

At Verifi, I have led the development of data-driven statistical and mathematical models that are translated into algorithms for the monitoring and control of the properties of concrete in a mixing truck during delivery. Streaming signals from multiple sensors are filtered and fused. Machine learning algorithms trained on historical data run in real-time on the truck and provide accurate predictions of rheological properties of the concrete in the mixing drum. I also have developed methodology for extracting actionable analytics based on archived data from many thousands of concrete deliveries; various metrics are shared with Verifi customers in monthly reports that provide visibility into the performance of the Verifi system, as well as the efficiency of operations at customer plants.Verifi® brings sophisticated, computer-automated manufacturing process control technology to the concrete truck. The Verifi system accurately measures, adjusts and documents concrete properties from the moment the batch is loaded to the moment it is placed.

As a Technical Staff Member in the Weather Sensing Group, I worked on several projects associated with the improvement of efficiency of operations in aviation. Examples IncludeTower Flight Data Manager (TFDM) is the next generation air traffic control tower information system that integrates surveillance, flight data, and other sources, which enables advanced decision support tools (DSTs) to improve departure and arrival efficiency and reduce fuel burn at the airport. As the leader of the TFDM DST algorithm team, I developed machine learning-based algorithms for the prediction of aircraft taxi times and landing times. In addition mixed integer linear programming methods were investigated for the optimization of aircraft departure and landing sequences.Consolidated Storm Prediction for Aviation (CoSPA) 0-8 Hour Convective Weather Forecast Using the Airspace Flow Program Blockage-based Capacity Forecast: In this project I developed methods based on Random Forests and Quantile Regression Forests for forecasting airspace blockage due to inclement weather up to 6 hours into the future.

Taught undergraduate classes in mathematics and statistics; directed undergraduate research projects; conducted original research in applied mathematics and statistics; published papers in peer-reviewed journals; presented research at professional conferences.

As a postdoctoral fellow and later as a visiting researcher, I conducted research in applied mathematics, with focus on fluid dynamics, turbulence, mathematical epidemiology and complex networks, published papers in peer-reviewed journals and presented research results at professional conferences.

As part of an interdisciplinary team of scientists at DTI, I led the development of mathematical models, based on ordinary and stochastic differential equations, of the spread and control of infectious diseases. The work was sponsored by the Armed Forces Medical Intelligence Center, and other defense and intelligence agencies.