From Higgs to Claude

A physics-to-AI career

Chris Arnesen · SJND Career Day · April 17, 2026

The early years — Minnesota

International School of Minnesota · ~10 kids in my graduating class

Math League · Show Choir · Sports

Turbo nerd who was also into basketball

The whole thing on one slide

Where	What
Minnesota	High school
Caltech	Physics + basketball
Minnesota	Gap year — coaching
MIT	Physics PhD
Carnegie Mellon	Higgs boson postdoc
Ab Initio	First software job
alwaysAI	Computer vision startup
Flock Freight	Cloud infrastructure
Tradeweb	Finance-tech (current) + personal-os on the side

Physics (and basketball)

Caltech · B.S. Physics (1997–2001)

4-year starter, men's basketball · team captain · Defensive Player of the Year 2000

Gap year — back in Minnesota

Coached 5th/6th grade boys basketball at my old high school

MIT · Ph.D. Physics (2002–2007)

Thesis: Model-independent approaches to QCD and B decays

Carnegie Mellon · Postdoc (2007–2009)

Calculated Higgs boson production rates at the Large Hadron Collider

I went to Caltech for undergrad. I studied physics because I was good at math and physics felt the most interesting — the most real, somehow. While I was there I played four years of intercollegiate basketball. Caltech's a Division III school, not a basketball powerhouse, but I was a starter all four years and captain my senior year. Basketball was probably as important to me as the physics, honestly. After I graduated I was burned out on school and wasn't sure about grad school. So I took a gap year, went home to Minnesota, and coached 5th- and 6th-grade boys basketball at my old high school. Low stakes, fun, and it gave me time to decide I did want to do a PhD. I went to MIT. My thesis was called "Model-independent approaches to QCD and B decays." Here's what that actually means. At particle accelerators, physicists are trying to learn about things like the Higgs, or new physics. But when particles collide, the strong force — the thing that holds the insides of particles together — is always in the middle of everything. It's messy and really hard to calculate. So if you want to get a clean answer about the interesting physics, you have to somehow see through the strong-force mess. My thesis was about mathematical techniques to do that. "Model-independent" means the answer doesn't depend on the messy details you can't get right anyway. Short version: clever math to pull signal out of noise. Then I did a postdoc at Carnegie Mellon, calculating how often Higgs bosons are produced at the Large Hadron Collider — the big particle accelerator at CERN in Switzerland. You may have heard it's where the Higgs was discovered in 2012. Cool work. It was also the moment I started realizing I don't actually love doing physics research day-in, day-out.

The pivot

2009 — left physics for software

Why:

I didn't actually enjoy research day-to-day — too tedious
The academic job market was brutal
I wanted to build things people actually use

By the end of my postdoc at CMU, I was starting to be honest with myself about something. I liked physics in the abstract — the ideas, the problems, learning it in college. What I didn't love was the day-to-day of being a physicist. A lot of grad school and a postdoc is staring at one calculation for six months. For some people that's heaven. For me it was tedious. Around the same time, the academic job market in physics was — and still is — brutal. There are way more people with physics PhDs than there are professor jobs. I was looking at the path forward and realizing: even if I wanted this life, I might not get it. And I wanted to build things people would actually use. Research has a very long feedback loop — you spend years on something maybe 50 people read. So in 2009 I left research and took a software job at a company called Ab Initio. More on that in a second.

Learning to code on the job

Ab Initio Software · 2009–2013

Software consultant — data processing for big enterprises

Zero formal CS training
Wrote software for banks, insurance companies, financial firms
Invented a new data differencing algorithm along the way

Ab Initio is a software company that builds data-processing tools. They hire a lot of physicists and mathematicians and don't particularly care if you have a CS degree. Which was lucky, because I had zero formal programming classes on my resume. I had a physics PhD and enough curiosity to be dangerous. What Ab Initio does: they sell really fast, reliable software for moving and transforming huge amounts of data — think the overnight batch processing that banks and insurance companies run. I spent four years embedded with those customers, building applications on top of their platform. By the end I had invented a new algorithm for computing differences between versions of big datasets. I hadn't known you could just do that — invent things because the problem in front of you didn't have a good solution yet. Two things I learned there that are still true for me. One: my physics training transferred directly. The PhD taught me how to stare at something confusing, break it down, and keep going until it made sense. That muscle is the single most useful skill I have. Two: most of what makes someone good at a job isn't what they formally studied. It's what they chose to dig into after they showed up.

Startup chapter

alwaysAI · 2019

Senior web developer — computer vision SaaS

Tiny team, early-stage startup
Designed & built the SaaS platform and CLI from scratch
Moved fast, wore every hat, shipped things real users touched

Somewhere in the middle of my software career I worked at a fledgling startup called alwaysAI. They were a small computer-vision company — their product helped developers run AI models on small devices like cameras and edge computers. This was pre-LLM-boom AI; the old-school neural-net era. What I want to tell you about startups isn't the technology, it's what working at one is like. I joined when the team was tiny. I designed and built their main SaaS platform and command-line interface from scratch. There was no existing codebase to inherit, no senior architect telling me what to do. I made a lot of the calls, some of them were wrong, and then I fixed them. I wore every hat. Every week something I had written the week before was being used by a real customer. Startups are a specific kind of intense. They're not for everybody. They're also the best place to learn at high speed, because the feedback loop is days, not years. If you like moving fast and are OK with some mess, a startup is a great place to get very good very quickly. If you like stability and well-defined roles, a bigger company is a better fit. Both are legit.

What I'm doing now

Tradeweb · 2020–now

Director, Technology Development — GraphQL APIs for anonymous digital financial exchanges

St. Paul Academy · 2021–2023

Head coach, middle school girls basketball (two seasons)

personal-os — AI + agent tooling side project

Claude Code + custom skills to automate my own life · [live demo]

These days I have three threads running at once. The day job is Tradeweb. I'm a Director of Technology Development. Tradeweb runs financial exchanges — specifically, anonymous exchanges where big financial institutions trade with each other without knowing who's on the other side. I build the APIs that connect people and software to those exchanges. Node.js, TypeScript, infrastructure. In parallel, I coached middle-school girls' basketball at St. Paul Academy for two seasons, 2021–23. Sixth grade team and a combined seventh-eighth grade team. Coaching keeps me honest about what I'm good at and what I actually like doing. And there's this thing I call personal-os. It's a big repository of code I've built to automate my own life using AI agents. I use a tool called Claude Code — I can tell it to check my email, pull my finances, book appointments, write code for me. [Live demo: show Claude Code running a skill.] The wild thing about right now, in 2026, is that what I just showed you would have been science fiction five years ago. And it's getting better every month.

3 things I wish I'd known in HS

1. It's not just what you know — it's who you know.

Connections compound. Grades fade fast.

2. Loving the subject isn't the same as loving the job.

Ask: what does this look like as a day job?

3. It's OK to be practical about your major.

You can love what you study AND let "what jobs does this lead to?" factor in.

Three things I wish someone had sat me down and told me at your age. These aren't universal — they're specific to me, and maybe they'll be useful to some of you. Number one: it's not just what you know, it's who you know. I spent college and grad school with my head down on my GPA. The people around me — classmates, professors, labmates — would have been the most valuable thing to invest in, and I mostly didn't. Connections show up in ways you don't expect: the next job, the promotion, the raise, hearing about opportunities before they're posted. Grades fade fast. Networks compound. Invest in the people, not just the grades. Number two: loving the subject isn't the same as loving the job. I picked physics because I loved the class. What I didn't understand is that being a physicist — even a professor at a liberal arts college — is mostly doing research, which I eventually learned I find tedious. I don't regret the PhD. I do wish I'd asked, earlier: what does this look like as a day job? Number three: it's OK to be practical about your major. I was idealistic at 17 and picked my favorite subject with almost no thought about what came after. Physics worked out for me in the end. But honestly — a CS major and going straight to industry would have been a better fit. You can love what you study AND let "what jobs does this lead to, and would I actually like them?" factor in. That's not selling out; it's being awake.

Thanks

Questions?

carnesen.com