Victor Jiao

About

Three Volumes · The Longer Version

Queens, Stuyvesant, and finding a shape

Vol. I opens in Queens, where I was born and raised, a child of immigrants. The early years were normal kid stuff (basketball, singing, learning English) before I tested into Stuyvesant.

At Stuy I kept singing, traded basketball for fencing, did math competitions (USAJMO, AIME, USAPhO), and co-organized hackathons with CodeDay NYC. I took the computer graphics and systems classes on offer, and taught myself C++, web dev, and AutoCAD in my own time, partly for projects I had in mind, partly just for fun. I liked balancing the technical side of my brain with the more artistic, more people-oriented one. I'd wanted to be a doctor, then a novelist, then (at Stuy) an architect, a game developer, an AI researcher.

The late night I remember most clearly is the one where I got Conway's Game of Life stable. Making it look alive was easy; keeping state clean between ticks, so generations stayed correct, was the trickier part. That insight stuck.

I found the shape of the person I'd become in that mix: someone who liked being specific, liked being part of groups, and liked the feeling of getting really good at something you could share.

UChicago, a cappella, and getting into graphics

Vol. II opens at the University of Chicago, where I did a four-year joint BS + MS in computer science.

In orientation week I caught Voices in Your Head at the a cappella showcase. Packed hall, dead silence when they started, and a set so tight I was trying to figure out how they made it look that easy. Standing ovation at the end. I knew I wanted in. I auditioned, got in, and ended up going to the ICCA Finals three times with them. Did ICPC with friends too.

I took a lot of hard classes. Briefly tried to triple-major in geoscience, physics, and CS because I'd convinced myself that climate change and AI were the two biggest threats to humanity. Dropped geoscience when the intro course turned out not to be calculus-based and went full CS. (Fun fact: I took graduate Climate Dynamics as a senior elective later on, and that one very definitely was more than calculus-based. It kicked my ass.)

UChicago's "take whatever you want" culture was real. I ended up in biodiversity and ecology instead of core bio, squeezed Electronics alongside Intro Physics III, and talked my way into Pattern Recognition, a second-year grad course, as a second-year undergrad with zero of the six prereqs. I hung in until the very end and then withdrew. I wasn't out of my depth because I was dumb; I just didn't have the statistical-theory grounding. Multivariate Gaussians, probability density and expectation, proving things via Bayesian inference and likelihoods. A lot to take in cold. Got an A in PhD machine learning the next quarter, which was a relief. And those hours half-grasping material laid surprisingly useful groundwork: semantic collapse in high-dimensional space, data curving into lower-dimensional manifolds. I think about those concepts at work now on AI memory.

A solid humanities core (Western scientific institutions, philosophy) plus a year of psychology gave the quantitative side a counterweight. Internships filled in the production side: Google on shadow maps for a 2D engine, Facebook inventing and integrating compression algorithms for a payment-fraud service. I liked how each balanced R&D with real productionization.

After graduating I moved west to Magic Leap. I'd loved their demos and wanted to see what the secrets were from the inside. Developed novel algorithms for image compositing there, which opened interviews at HoloLens and Daydream; I joined Google Stadia instead. Stadia is where I really started naming the thing I cared most about. My tech lead pulled me aside in our first 1:1 and asked what I actually cared about in software. I thought for a second and said: that moment of magic. When you play a game and something happens, or a UI clicks into place, or a simulation surprises you. I wanted to make those feelings, anywhere up and down the stack. She smiled and said she thought I'd have a chance to really explore that there.

SF, community, and Meta Reality Labs

Vol. III opens with a move to San Francisco, just as the pandemic was starting to lift. The first stretch was a lot of community-finding: playing volleyball with NAVGA, co-founding The Chromeatics at Google SFO, helping re-start Pride@Google-SFO for a while. I spent a lot of it watching how different groups worked: how leadership formed, what structures they leaned on, how they held together.

Stadia got cancelled. I landed back in VR, first at Google Labs and then at Meta Reality Labs as senior SWE on the input & interactions team. That was a real jump, and a genuinely cool one. I can't talk about the specifics of what we're building yet, but I love the work: sitting between prototypers and designers, shaping interactions for new VR systems, productionizing the ones that worked, and bringing Stadia-era frame-rate discipline into the pipeline.

When AI started getting actually good, I led my team in figuring out how to fold those gains into engineering: new skills, usage patterns worth advocating for, and memory systems for individual engineers and for teams. Some of what I was building by the end of 2025 is now popularized and mainstream. I'm still happily balancing the two.

These days I sing with the SF Gay Men's Choir, throw pottery and do a bit of woodworking, and host AI hack nights at home. The throughline, as far as I can tell from inside it, is that I'm spending more time now on what seems to matter most: finding and building community, and trying to grow as a human.

If you want to reach out, email is best. The résumé-shaped version is here.