Math
Before I moved to Computer Science, I was a math PhD student for three years. I was passionate about the Langlands program and was researching things on the Galois theory side and on the Lie theory side at the same time. I had a handful of projects; In Galois theory, I computationally verified Buhler's thesis (on modularity of Icosehedral Galois representations). And I was researching what corresponding things could be said about even Galois representations. On the Lie theory side, I started by learning about infinite dimensional representations of real reductive Lie groups, and worked my way up to Automorphic representations of $GL(n)$. Then I learned about Cohomological Parabolic Induction, I was starting research on these cohomology groups before I left. I often think if one small thing was different, I would have stayed in math. Writing about this makes me miss it, I get tempted to go back and learn this little point, and that little point. This area is very beautiful. I don't think anything compares!
I might write more on this later, I could also share the specifics of my work to anyone interested. This website only contains my work in CS. Here's a CV containing my work in math.