• Developed and integrated machine learning and computer vision algorithms for performing pixel-level aerial image matching, improving accuracy from approximately 75% to 99.9%• Designed and constructed a distributed network of computing and monitoring systems to implement high volume image processing and analysis pipelines• Constructed a framework for using quantitative metrics and statistics to assess algorithm performance and improvement• Researched and implemented novel deep convolutional networks for image segmentation and pixel-level image matching

Department of Genetics under Dr. Lisa Strug• Harnessed machine learning to build a near clinical-grade tool for predicting the likelihood of comorbidities such as meconium ileus and decreased lung function in patients with cystic fibrosis based on genetic data• Built and compared the predictive power of models including deep neural nets, random forests and LASSO regression• Analyzed the genetic information used by the model to identify biological motivations that could improve the model

Department of Biochemistry under Dr. Hue Sun Chan and Dr. Lewis Kay• Optimized a C++ model to carry out Monte Carlo simulations of interactions among polymers• Devised efficient algorithmic solutions to complex logical and three-dimensional problems related to protein structure, movements and interactions• Validated an analytic theory regarding polymer end-to-end distance using simulated data points• Created plots and visual representations of data output in MATLAB to convey key insights