During my B.Sc. (Honours) in Physics at the University of Guelph, I developed a passion for learning how the world works from a mathematical perspective. In my 4th year, I completed two thesis projects, one was in physics and the other in math. The physics project involved calculating the location of the neutrino surface corresponding to a simulated accretion disk and for the math project I explored an inverse problem, namely, to estimate the parameters of a fractal given its image. The project offered a solution using a genetic algorithm with a custom fractal image metric that I had devised.My enjoyment of the math project led me to continue on with an M.Sc. in Math in which I developed a thesis involving two components: I attempted to improve the results of my 4th year project using CNNs, and to generalize theory on iterated function systems, where I allowed them to include other function classes and developed fractal splicing.After completing the M.Sc. I came to UBC to further develop my knowledge of machine learning. My thesis project is about estimating specific user traits via eye-tracking and other interaction data to personalize, in real-time, the tutoring system to the user while they work on problems. I tested two types of feature sets to accomplish this task: summative statistics of the data (e.g. the number of fixations and saccades), and the frequency of behaviours found through data mining. A statistical comparison of both feature sets along with examination of feature permutation importance led to finding important insights as to why some features were better at predicting one trait over another.Check out my GitHub at https://github.com/broccoloni to see some of my projects!