• Developed metrics to predict top submissions based on user preferences

• Used natural language processing and deep learning to further improve the models based on the content of the submission

• Created a web app to aid in the discovery of top submissions before they become popular

• Built an interactive web app to fit probability distributions to input data and

graph the results

• Interactive display and plotting allows users to visualize different models and

parameter values and save the values of the best fit

• Created an interface to analyze the relationship between any set of properties in

Lumiata’s medical graph using the model fitting tool

• Used this to determine properties of interest in the medical graph by seeking out

atypical results, such as insurance providers with higher incidence rates for

specific diseases

• Collected personal records and financial histories of NFL athletes using Python

• Analyzed athletes’ spending habits in order to investigate economic models of

intertemporal consumption

• Presented results at Caltech SURF Seminar Day in Oct. 2013