• Analyzed multi-product CONWIP system with complex interactions
collaboratively
• Used MVA approximation to find workstation and factory performance
measures with limited information
• Wrote all the code for calculations in Python
• Used Bayesian machine learning methods to predict umpire performance
on 3 different large data sets
• Compared and analyzed performance of various models to determine best
predictors
• Collaboratively worked on the report after I developed the models in
Python
• Built upon existing research provided by U.S. Army DEVCOM DAC
• Utilized academic resources and faculty to advance research progress
• Discovered and discussed multiple new failure modes applicable to
federated learning systems