About

I'm starting Computer Science at UC Santa Barbara in September 2026. I started programming when I was nine, in Scratch, and spent the next few years making Unity games and small websites that never went anywhere. Since December 2020 most of it has gone into one thing: a programming language and its compiler, an app deployment platform, and an attempt at turning all of that into a company.

The thing that runs through all of it is that I want to know what is underneath. Knowing that something works doesn't really satisfy me. I want to know why, and the way that reliably happens for me is building my own version and watching it fail in useful ways.

I learn by building. That's less of a philosophy than an admission, since reading about something doesn't stick for me and writing it does.

The loop is pretty consistent. Something looks interesting, I read the docs and then the source, I build a version of it, it breaks, I work out why, I fix it, and by then I understand it better than reading ever would have gotten me. Heap compaction in ProxOS is a good example. I understood why garbage collectors need to know which words are pointers about four minutes after my own compaction pass invalidated every reference in a register file.

A lot of that started on YouTube. Most of what I picked up about lower level programming came off courses on there rather than out of a class, and it is still where I go first when I want to understand something new.

Programming languages, mostly. Compilers, type systems, code generation, LLVM, and the tooling that has to exist around a language for anyone to use it. Pekoscript is where most of that went. Writing a type checker and then an LLVM backend for a language I designed taught me more about language design than reading about it ever did, mostly because it forced me to be precise about things I had been vague on.

Systems too. Operating systems, memory, binary formats, low-level programming. ProxOS is the small version of that.

Developer tools in general, which is what Peko turned out to mostly be. And the problem-finding side of entrepreneurship, which I used to think was the boring half.

I used to think the hard part was building. If you could build the difficult thing, you'd earned the outcome.

Peko settled that pretty conclusively. I did build the difficult thing. The two-pipeline split is still the best engineering decision I've made and it was worth nothing commercially, because I'd never established that anyone needed it enough to switch.

What I think now is that how hard a problem is tells you nothing about whether it's worth solving, and that being willing to stop is a real skill instead of a failure. That's why the loan processor thing is research and not a startup.

I did football, wrestling, and distance track my freshman year of high school, playing defensive end in the fall. Wrestling is the one that stuck. Four years of it, two varsity, captain the last two seasons. These days I train jiu-jitsu, which I'd originally done in middle school before wrestling took over.

I spent years in Scouting and ended up as Lodge Chief of Tiwahe Lodge. My Eagle project was a veterans memorial wall that I started planning in eighth grade, and it's the first genuinely complicated thing I had to organize other people to pull off.

And I listen to a lot of R&B and yacht rock, which needs no justification.

I don't have a clean ten-year plan and I'm suspicious of people my age who do.

What I know is that I want to be independent, get really good at hard things, and build stuff that actually matters. I'm more interested in finding a problem worth going deep on than in deciding early what my job title should be. Depth over breadth, basically. I'd rather understand one hard thing completely than forty things approximately.