✍️ Musings

Impermanence to purpose

At twelve, I learned how abruptly life can change when someone I loved disappeared from my daily orbit. That shock didn’t make me cynical; it made me restless for answers. I became consumed with the question of how we could sustain longer, healthier lives, and technology felt like the most audacious tool available. Ever since, I’ve treated every topic as a puzzle to be disassembled, interrogating first principles and asking the extra question that people usually leave unspoken.

When curiosity outruns exams

In Chinese high school, everything revolved around the gaokao (高考). I never cracked the top scores, but I was the student haunting the teachers’ office, asking my physics teacher why gravity feels like an invisible glue binding particles. That instinct to chase first principles — instead of test hacks — soaked up time and showed in my scores, even as teachers called me the hardest worker in the class.

With more practice I learned to play the exam game, yet I never matched the naturals who vaulted into the most resource-rich colleges. The system’s promise of equal opportunity matters, but it can flatten curiosity in the process. I still hope we can keep fairness while making more room for students who can’t stop asking why.

Hacking systems beyond code

During undergrad, I spent seven months building a bioinformatics platform by myself — no advisor, no syllabus, just a stubborn curiosity. I taught myself Java, protein-network algorithms, and enough computational biology to launch BNMatch2, an open tool that would later inspire the semantics modeling in my first top-tier CV publication. The real breakthrough wasn’t the code; it was the realization that insights from protein information flow could be smuggled into AI, proving disciplines talk to each other if we force the conversation.

A moment that reshaped my path

One midday in college, I climbed on stage and delivered an English campaign speech while the entire campus milled around with trays of food. It was audacious and a little chaotic — and it worked. That moment led to my selection as one of thirty student representatives for a leadership program at The Wharton School, validating that ideas travel further when you’re willing to voice them in imperfect settings.

Choosing conviction over certainty

After graduation, I declined a guaranteed spot in graduate school. Family and friends questioned me, but I packed for Shenzhen anyway to work as a research assistant at CUHK(SZ). I lived with the uncertainty, pushed through, and authored my first research paper before many senior lab members published theirs. That experience rewired how I evaluate risk: I now trust conviction more than consensus.

From papers to products with AI

During my PhD at HKUST(GZ), I kept seeing research constrained by narrow setups, resource-heavy experiments, and results that cracked under real-world pressure. Watching tools like Cursor automate coding made the pattern clear: in fast-moving fields like AI, a paper can be leapfrogged within months. Spending half a year for one top-tier publication is an achievement, but the next release often eclipses it.

I came to believe most of that work is 90% engineering polish that automation will absorb. The opportunity is in the remaining 10% — taste, design, and judgment — and that’s where business lives. I spent the next half year building HoloSoul from scratch as a real product, not a PDF: system software aimed at customers, revenue, and durability. That pivot convinced me I’d rather build companies that survive contact with the world than chase citations that don’t.

What I’m known for

Colleagues describe me as someone who disassembles complex systems until only the critical gears remain. I lean on intuition, creativity, and first-principles reasoning to hunt for the leverage point — the single insight that shifts the entire problem. That reputation keeps me accountable to the craft of deep thinking.