Research

I am broadly interested in programming languages, compilers, verification and concurrency. I’m particularly motivated by methods that empower programmers to write correct and efficient software in the real world, bridging theory and practice. For my master’s thesis, I worked with Maximillian Doré on designing and implementing a new programming language that combines algebraic effect types and quantitative types using a behavioural effect type system. For my bachelor’s thesis, I worked with Nobuko Yoshida and Joe Paulus on developing the syntax, semantics and model-checking procedure for probabilistic multiparty session types. I was awarded the Department of Computer Science Prize for Best Project in Part B for this work. I’ve also collaborated with Billy Moses, Albert Cohen, Alex Zinenko and others on integrating equality saturation into Google’s XLA machine learning compiler.

About Tyme: Reconciling Algebraic Effects and Linearity Through Behavioural Effect Types
Leo Seojun Lee
Master's thesis
Verification of Probabilistic Distributed Protocols with Multiparty Session Types
Leo Seojun Lee
Bachelor's thesis
Best bachelor's thesis prize
Mind the Abstraction Gap: Bringing Equality Saturation to Real-World ML Compilers
Arya Vohra*, Leo Seojun Lee*, Jakub Bachurski, Oleksandr Zinenko, Phitchaya Mangpo Phothilimthana, Albert Cohen, William S. Moses
OOPSLA '25