Ploy Phusri
This summer I did applied machine learning research at UBC, combining topological data analysis with convolutional autoencoders to classify cell images. Before CS, I worked full-time in UX/UI design, management consulting and sales engineering.
what i bring
I get up to speed on hard technical topics fast.
This summer I learned persistent homology and convolutional autoencoders well enough to build a research pipeline that turns cell images into features. Before that, I picked up zero-knowledge proofs to build zkDebit.
- ML/TDA research · summer 2026
- Food classifier · random forest
- zkDebit · zero-knowledge proofs
I make complex ideas legible.
I find the structure inside dense material and explain it clearly: in course notes, in essays, and in a platform that teaches deep topics through visualization and makes them fun.
- ployground · interactive learning
- Thoughtstack essays
- CPSC 110 notes
I build for the people who use it.
Working in consulting, sales engineering and UX design, and founding my own startup, taught me to see a product through the eyes of its users, its designers and the business.
- Celatone UX internship
- Blockbook · founding a startup
- Experience · LinkedIn
selected work
view all →TDA + autoencoders for cell classification
Persistent homology and convolutional autoencoders turn cell images into features a classifier can learn from.
MonsterTamer
A personal emotional RPG in Java that turns anxiety, procrastination and bad habits into monsters you tame with coping strategies.
zkDebit
Built at an Ethereum Foundation zero-knowledge hacker house: shoppers prove they hold a valid card instead of handing their card number to every merchant.
building on the side
ployground
An interactive learning platform for deep topics.
recent writings
view all →current obsessions
Outside of school, I teach yoga at UBC Yoga Club, play Go, and sing and play piano. I also help run UBC AI Safety.