United Airlines (06/2009-08/2011)
Senior Analyst, Finance-Operations Research       Chicago, IL
New forecasting and optimization algorithms for next generation revenue management system and simulation analysis of different designs.
Development of important components of a network revenue management simulator.
Mileage redemption seat displacement cost analysis.
Forecast quality evaluation via historical demand tracking.
Discount full fare class management study.
Analysis on the impact of real-time inventory availability on forecasting and inventory optimization.

IBM T.J. Watson Research Center (06/2008-09/2008)
Business analytics and mathematical sciences intern     Yorktown Heights, NY
Designed importance sampling method to reduce variance of a simulation based Conditional Value-at-Risk estimator used in a stochastic optimization framework with chance constraints.
Implemented a sampling procedure to make random draws from a random vector with any specified marginal distributions and correlation structure.

United Airlines (06/2007-10/2007)
Enterprise optimization intern      Chicago, IL
Enhanced market level discrete choice model based forecasting procedure and reduced forecasting error by 65%. The algorithm was made up of a linear programming based spill and recapture procedure for flight capacity allocation and a nonlinear simplex optimization algorithm for market size calibration.

IBM China Research Lab (06/2005-09/2005)
Business optimization intern        Beijing, China
Formulated a multi-echelon multi-commodity supply chain restructuring problem.
Designed and implemented a hybrid genetic algorithm and linear programming supply chain optimization solver and collaborated with system developers to interface the solver with geographical information system.

Shanghai Optical Networking Tech. (05/2000-10/2001)
R&D Engineer, Co-founder           Shanghai, China

Designed and implemented a wavelength monitoring unit for a Wavelength-Division-Multiplexing ring network.
Designed a hybrid genetic algorithm for traffic grooming to reduce electronic multiplexing.
Conducted comparative studies of the performance of simulated annealing, tabu search, and genetic algorithms for traffic grooming.