I build and run a fleet of AI agents across five machines, and I point it at things people usually hire a team for: quant trading, a few thousand crypto nodes, and a token allocation I earned by actually trading on the platform. It's all one person.
Every claim here is tagged so you can check it yourself: live, on-chain, or historical.
On-chain proof
The tokens landed straight in the wallets that qualified, based on how much I actually traded. There was nothing to claim and no vesting schedule.
of Lighter's first airdrop, a 250M-LIT pool paid on-chain to the wallets that qualified.
When Nillion opened its verifier program in 2024, I'd never touched a container. I learned Docker with AI as I went, and had the whole fleet running myself.
verifier nodes in isolated Docker containers, on a Mac mini cluster I run at home.
It's the legal home for this work. I set the whole thing up myself: the domain, the company email, the GitHub org, and the automations that keep it active day to day.
the age I registered it. Taiwan entity live; Hong Kong entity in formation.
I designed the whole system myself: the isolation between environments, the state machines that tracked each one, and the batch logic that drove them all at once. Hundreds of environments, all behaving the same way, with no one watching.
HistoricalFive machines running agents around the clock. When one freezes or crashes, the system notices and brings it back with no one watching. I built the orchestration and the recovery around the models. I didn't train a model.
production failures turned into automatic recovery. One silent freeze that ran for hours now recovers on its own in about 100 seconds.
Quant systems
A cross-exchange latency-arbitrage engine, written in Go. I did the strategy, the signal work, and the execution engine myself, all on my own Hermes fleet. The strategy didn't come from a book. I reverse-engineered it from public fill data on the exchanges.
exchanges wired into one Go latency-arb engine, running live on my own fleet.
Origin
I went to 師大附中, one of Taiwan's top high schools. I taught myself most of the coursework with AI instead of sitting through it. I took a gap year and left two credits short of a diploma, because re-taking classes to hold onto a piece of paper was not worth the time. So I have a certificate of studies, not a graduation certificate.
The gap year is when I actually started building. I was already deep in crypto, running my own automation at scale. Once the coding models got good in 2024, I leaned on them hard: I taught myself Docker from zero and stood up a Nillion verifier fleet on my own, past 5,000 nodes.
Then I got into 清華大學 (NTHU) through the 拾穗 program, an admissions track that ignores standardized test scores and looks at your work and an interview.
They admitted me on my work, not a test score.
It's the same thing I did in high school, just bigger: use AI to build whatever I want, and don't half-do it. The agent fleet, the quant systems, the company all came out of that.
The Academy
The way I work is AI-first: one person plus a fleet of agents, and crypto is where I put it to the test. On-chain there's no hiding. The trade either happened or it didn't, and the allocation either landed or it didn't. I've been working this way since high school.
Of the founding partners, my work lines up with Coinbase most directly. I've already built exchange systems, execution engines, and automation around on-chain incentives. I'm not looking to explore crypto as an application. I've been shipping it, with on-chain proof to show for it. Give me the $50,000+ in partner compute and my edge is turning it into systems that do the work of a whole team.
Most people here will be strong at one thing. I've got three that build on each other: the agent fleet I engineered, the allocation I earned on-chain, and the quant system running on top of both. All of it solo.
And I'm all in. If you take me, I leave 清華大學, move to San Francisco, and spend the year on this full-time. I've bet on work over credentials before, when I walked away from my diploma. I'll do it again.