Growing up, I always enjoyed math. I initially focused my time on sports (hockey and bodybuilding), but once I realized that I had more talent for math, I leaned into it hard. In high school I self-studied MIT's undergraduate math/physics curriculum, completed two research projects at local IUSB and Notre Dame physics labs that improved acoustic and optical data transmission in particle detectors, worked as a Mathnasium instructor on evenings and weekends, wrote some expository material, and derived a nasty closed-form expression for decomposing a general rational expression into partial fractions. I wanted to create mathematical models to describe all physical, biological, economic, and social phenomena, starting with the brain (since understanding and improving human intelligence could speed up progress in other areas as well).


In 2014, I attended Notre Dame on a full-ride Lilly Scholarship and majored in math. I ended up in a handful of 400-level and grad classes as an underclassman and had a couple professors take me under their wing with expository projects, but I quickly became disillusioned with the general lack of relevance to my modeling goals (the sole exception being a project for game theory class), so I distracted myself with independent projects. After becoming the bajillionth student to burn time investigating the 3n+1 problem, I embarked on a quest to model the human brain as a dynamic weighted graph. My initial approach was to abstract from first principles, and I began studying how the topology of biological neural networks with spike-timing dependent plasticity can (in theory) be shaped by periodic stimulation. While this led to an interesting toy problem that was tractable in a simple case, the results offered little intuition and the complexity ramped up very quickly, so I pivoted to a top-down approach. I interned with the synthetic cognition group at Los Alamos National Lab the summer after my freshman year, attempting to create an emergent phenomenon of brain oscillations by implementing spiking neurons in a deep learning model. It turned out to be an incredibly difficult task and, beyond implementing a form of spiking neuron, the project wasn't very fruitful (HN comments here and here are relatable). On top of the sheer modeling complexity, there was also a lack of experimental data on the ~100 billion individual neurons and ~100 trillion synapses in the human brain, and I determined that developing a satisfactory model of the brain was well beyond the scope of a single lifetime (for me, at least).

In search of more tractable, practical, and financially rewarding modeling problems, I jumped to industry and interned as a data scientist at Aunalytics during my sophomore year. The executives were impressed by my contributions, so I worked full-time throughout my junior and senior years while simultaneously finishing my degree. My projects focused on churn prediction and exploratory analyses in financial services and digital news, as well as evaluating the potential of topological data analysis (which turned out to be elegant in theory but not so useful in practice). I gradually realized that quantitative modeling was far more valuable when paired with domain expertise, and although I was grateful to have broken into data science so early, I wasn't excited by any of the fields that were conventionally paired with data science. Instead, I was interested in math education and music production (my main hobbies were making music and continuing to work as a Mathnasium instructor on evenings and weekends). I had more talent for math education, so after graduating from Notre Dame in 2018, I left Aunalytics to throw myself fully into math education and gain more domain expertise while earning a master's in computer science from Georgia Tech (remotely). I also moved to California to be close to my girlfriend Sanjana at Caltech.

Math Academy

Initially, I took on scattered freelance work in math education: tutoring for many different agencies, developing content for several online learning platforms, and teaching high school and weekend test prep. I also wrote several textbooks for fun. Gradually, all my tutoring/teaching and content development converged at the same organization, Math Academy, a highly accelerated 6-12th grade math program where 8th graders take AP Calculus BC and high schoolers study a full undergradute math curriculum. Math Academy ran on top of educational software built by the founder, Jason Roberts, a serial entrepreneur who, after developing much of Uber's foundational technology, wanted to create and commercialize the ultimate online math learning system. I got involved at the core of Math Academy's product during the summer of 2019, when Jason hired me to develop an algorithm that would automatically create personalized assignments while leveraging effective learning techniques like spaced repetition and interleaving. By the end of summer, the project was a success, and it upgraded the software from a teaching tool to a fully automated and personalized learning system that could effectively support independent learners (without any teacher).

Since then, I've spent nearly every waking hour involved in all aspects of the product, with a focus on algorithm development. Most notably, I created and continue to improve our Fractional Implicit Repetition (FIRe) model, which generalizes discrete spaced repetition on independent tasks to fractional implicit spaced repetition on highly connected knowledge graphs and is orders of magnitude more computationally efficient than conventional algorithms on knowledge spaces. We're currently beta testing!

*           *           *

Thank you to the following people who have significantly impacted my life:
  • My parents - for always encouraging and wishing the best for me
  • Ken Andrzejewski and Douglas Sisk - for teaching me how to put myself out there, find research opportunities, and compete in science fairs during high school
  • Ilan Levine - for advising my first high school physics research project
  • Cari Ingram - for hiring me as a Mathnasium instructor during high school and letting me have so much fun goofing around with the kids
  • Dave Cieslak - for hiring me as a data scientist during college and sending interesting projects my way
  • Jason Roberts - for helping me find my niche after college and providing a constant stream of challenge and mentorship