My Journey from Traditional Computer Science to DeFi

It's been just over a month since I started my first job, and I got some great news: my manager told me I've passed the probation period and am now a permanent employee! I work at a DeFi start-up in Hong Kong as a Defi Quantitative Developer with about 20 people, all super nice and supportive.

Honestly, I never thought I'd end up in DeFi, blockchain, or anything Web3-related. Me, I'm a "traditional" computer science guy at heart—compilers, operating systems, that's my jam. I even wrote a bunch of blogs about this stuff back in the day. Here are some of my favorites:

• Parsing Techniques
• Writing a Small C Compiler
• Virtual Memory
• OS Interrupts and Context Switching
• Special Topic: Traps on Xv6
• File System I & II

Those were pure, golden times: Being 20, at university, free to explore whatever I wanted without a job tying me down.

Then came COVID-19. I'd gone back to Shanghai from Hong Kong, and like everyone else, I had to quarantine in a hotel for 14 days to stop the spread. On day seven, I tested positive for COVID and got moved to a hospital. I ended up in a room with four other people who also had it—kinda like a jail, but worse. You're lying there, surrounded by sick people, not knowing when you’ll get better or get out.

It'a weird feeling, you know, like being trapped in this environment full of the virus with no escape. To keep my sanity, I started MIT's 6.S081 Operating Systems course online—thank goodness for schools like MIT uploading lectures to YouTube during the pandemic. The professors, Frans Kaashoek and Robert Morris, were amazing. I watched every lecture, did every lab, and took tons of notes.

One night stands out: I was wrestling with the Xv6 OS trap mechanism—how it jumps from user state to kernel state and back via this thing called the trampoline. I was lying in bed, struggling, while the old man next to me listened to novels on his phone (he could barely see without glasses) and another guy scrolled TikTok all night. After hours of thinking, I finally got it and jotted it down in my notes: Traps in Xv6. That moment felt huge—like I'd cracked something big.

That struggle taught me something: education, especially at university, isn't just about the knowledge itself. It's about the grind—learning how to learn, managing your time, pushing through when it's tough. Sure, I love CS, but the stuff we learn in school isn't some secret sauce only professors can teach. You could pick it up on your own if you wanted.

For me, the real value of those four years is bigger: it's about becoming someone who can tackle anything, and struggles/make mistakes in a controllable/saft enviroment–rather than society.

I love the idea that people spend time on physics, math, or philosophy—even if it doesn't directly "help" their job later—because those struggles, grindings makes you a better person, ready to grow in any field.

Back then, I pictured myself as a system software engineer working on databases or maybe a professor teaching others—I love explaining things. DeFi? Never crossed my mind.

After undergrad, I went to grad school thinking I'd dive deeper into traditional software like compilers or operating systems.

But in Hong Kong, there weren't much jobs choices for that (lol, I was so naive), and in grad school, it was the same—not much research on distributed systems or databases, just AI and LLMs everywhere. I kind of liked things I could explain—stuff with clear principles. That's why I avoided AI, even though it's popular and useful. To me, AI feels messy; Most people tweak parameters without knowing exactly why they work.

The only paper I could understand was about scaling blockchain. It felt like traditional distributed systems, so I wrote a report on it. My advisor liked it and agreed to supervise me. That’s how I became a grad student and research assistant in blockchain, focusing on layer 2 scaling solutions for the Ethereum Virtual Machine.

And soon I was hooked on ZK-Rollups. Compared to stuff like Plasma or Optimistic Rollups (which need a challenge period and don’t settle fast), ZK-Rollups have rapid finality and tons of potential use cases. I started dreaming big: fully scaled blockchain tech could replace centralized systems with decentralized, trustless networks. Smart contracts could automate everything—agreements, operations, you name it—on a global scale. Crypto's just the tip of the iceberg.

I got obsessed. To figure out ZK-Rollups, I dug into verifiable computing and wrote a survey: A Survey of Interactive Verifiable Computing.

It covers protocols like Sum-check and GKR, with proofs and all. Then, over New Year’s, I studied ZK-SNARKs—elliptic curves, QAP, pairings—and wrote a small paper called New Years Resolutions to propose my MPhil thesis on benchmarking ZK-Rollups. It was my way of showing my professor I got it.

In the end, I decided a PhD wasn't for me. I wanted to jump into the real world while finishing my benchmarking research for degree. That's how I landed at this DeFi company. The job's been great—not too hard, especially with AI breaking problems into bite-sized pieces for me. Since it's blockchain-related, it fits my vibe. I still work on my ZK-Rollups benchmarking during the day (when I'm done with tasks) or at night, fresh from a shower, chilling on my bed with my laptop.

Today after taking the shower, I've been thinking a lot about this journey. So much has happened—decisions I made, friends I lost, new ones I gained. Yesterday, an old friend I hadn't talked to in almost 10 years called, asking if he could visit me in Hong Kong. Crazy how life twists, right?

Looking back, I find myself praying a lot: "God, grant me the serenity to accept the things I cannot change, the courage to change the things I can, and the wisdom to know the difference.".