Geo Lee
Studying Computer Science and Mathematics at The University of British Columbia (UBC) in Vancouver, with interests in music and audio computing, machine learning, and computational approaches to musical creativity.
About
I am interested in research on musical creativity in AI, especially the problem of what a creative music model would require beyond surface-level stylistic imitation. I am particularly drawn to generative models for music and sound, representation learning for musical structure, and questions about how composition, production, and expressive decision-making might be modeled computationally. Outside formal coursework and research, I also produce electronic music independently, which is a major reason I am drawn to music and audio research, especially questions of musical structure, sound design, originality, and what creative AI systems for music might look like.
Experience
My recent experience combines machine learning systems work with audio-related pipelines, real-time experimentation, and research engineering. It has given me practical experience that I want to apply more directly to music and audio research.
Undergraduate Research Assistant
Worked as an undergraduate research assistant in Dr. Frank Wood's lab on real-time machine learning systems in a non-simulated interactive environment, including low-latency execution interfaces, multimodal state tracking, distributed experimentation, and evaluation under realistic system and latency constraints. I also built audio and transcription pipelines on AWS and contributed to logging, orchestration, and GPU-enabled inference infrastructure for large-scale experiments.
Undergraduate Teaching Assistant, CPSC 221
Led tutorial sections in data structures and algorithms, supporting algorithmic reasoning, implementation correctness, and complexity analysis.
Select Projects
Selected projects reflecting my interests in machine learning, music and audio computing, and mathematically grounded computational work.
A Comparative Analysis of Embodied AI Agents in Minecraft
Developed a live multiplayer experimental environment for evaluating agent execution beyond simulated benchmarks. Compared symbolic planning and reactive control strategies under real-time system constraints, and designed evaluation protocols measuring robustness, responsiveness, and task performance in a dynamic interactive setting. This project combined experimental system design with practical evaluation of embodied agents under real-world latency and coordination constraints.
Deep Neural Network Based Music Key Detection System
Designed and implemented an end-to-end machine learning pipeline for music key detection from audio recordings, including data collection and preprocessing, feature design, model training, and evaluation. Through this project, I became interested in the relationship between audio representations, prediction confidence, and the role of musical structure in learning-based systems.
A Software Simulation of the Proof of Roth's Theorem
Implemented a modular software simulation of ideas related to Roth's theorem using discrete Fourier analysis. The project explored computational perspectives on combinatorial proof structure and mathematical abstraction, and reflected my broader interest in formal reasoning, algorithmic structure, and mathematically grounded computational systems.
Select Awards
Selected awards and research funding reflecting academic performance, research potential, and support for independent research.
Advanced Machine Learning Training Network (AML-TN) Award
Selected to join UBC's Advanced Machine Learning Training Network, which provides research funding, mentorship, and mobility support for collaboration and training opportunities with leading institutions including MILA, the Vector Institute, Oxford, and MIT. Supervised by Dr. Frank Wood.
Science Undergraduate Research Experience (SURE) Award
Received funding for full-time summer research through the UBC Faculty of Science. This support contributed to the development of real-time transcription systems and infrastructure engineering for embodied AI data collection. Supervised by Dr. Frank Wood.
Natural Sciences and Engineering Research Council of Canada Undergraduate Student Research Award (NSERC USRA)
Received the NSERC USRA through the UBC Faculty of Mathematics in recognition of academic performance and research potential. Supported an undergraduate research project in combinatorics and incidence geometry. Supervised by Dr. Joshua Zahl.
Governor General's Award
Awarded by the Governor General of Canada to the graduating student with the highest academic average at a high school.
Contact
Links and materials.