Machine Learning & Data Science · PhD, Computer Science
I turn advanced mathematics into performant, well-tested software — machine learning, statistics, and reinforcement learning. I reduce a hard problem to its core, design experiments that build understanding, and ship correct, reproducible results. Python · Julia · HPC.
Turning dense, high-dimensional data into figures you can read as a pattern instead of a table of numbers. Built in Julia with Makie.
Art's hidden topology
First-author paper in PLOS Computational Biology: how the mathematics of shape explains human perception of abstract art. Open-access, journal cover, and covered by Scientific American. I ran the full statistical analysis.
MesoSCOUT & TDA packages
Open-source Julia framework that finds mesoscale structure in large networks from the relative strength of connections. First-author preprint, reproducible pipelines, published packages.
Empowerment for open-ended learning
An information-theoretic reinforcement-learning agent (Python, JAX) that decides when to discover a new goal versus master a known one. IMOL 2025 workshop; extended manuscript in preparation.
For the research-to-industry story, see About; for the full record, see Publications.
