projects
I tried teaching a Julia model to map one function to another, from integrals to PDEs.
přehled základních konceptů strojového učení
classifying protein sequences into their most common UniProt families using pretty basic NLP tricks. I tried k-NN, decision trees, and random forests and compared how they did.
I pulled together GDP figures across Europe and tried to make sense of the gaps between countries.
a small tool for exploring the bernstein-vazirani algorithm under hardware noise.
I tried to explain nitrate levels at river mouths, starting with population density and a simple regression.
I turned Prague's timetables into a graph to see which stops hold the network together.
how many of those 784 pixels do I really need to tell a T-shirt from a pair of trousers?
I compared crime counts across Austrian regions, then looked at what changes when you account for population.
a little maze where the right route matters as much as reaching the exit: each tile changes your quantum state.
does arm bone length change whether sparrows survive winter storms?
I dug through Austin's shelter records to see how animals arrive, leave, and sometimes come back.
predicting life expectancy from country-level health and economic indicators. Lots of feature engineering, then random forest, ridge, and k-NN pipelines compared head to head.
I kept changing the model to see what actually helped with Fashion MNIST: depth, convolutions, dropout, batch norm.
a robotic art piece about surveillance, inspired by George Orwell's 1984.