Henry Zweiman
I’m a student at Brown University, working on applied AI, mathematics, and bioinformatics. Mostly, I’m interested in what happens when you give AI a real problem, a way to experiment, and something useful to measure.
As a Kaggle competitions expert, I use AI-forward approaches to develop ideas, run experiments, and build competitive solutions. I build the systems around that process, with a focus on useful evaluation and results that hold up beyond a single experiment.
Lately, I’ve been working on well-known open problems in mathematics and ways to optimize LLM post-training. I’m also co-founder and CTO of UCeve, where we apply machine learning to develop treatments for rare genetic diseases.
On the biology side, I built Deoxyer, a tool for looking up sequences, optimizing codons, and assembling genetic constructs. I like making complicated research workflows easier to work with.
I also spend a lot of time on mathematics, using computation and AI to explore questions in algebra, number theory, analysis, and combinatorics. I’m especially interested in how agentic workflows can produce new mathematical arguments.
If you’re working on something difficult and think AI could help, I’d like to hear about it.