I build teams that ship.

I landed at Boltmade as the first non-founder—technically an intern. Five of us; four years later, nearly thirty. We built first prototypes for funded startups with no team, dropped into mid-sized companies stuck in politics, and ran skunkworks for Fortune 500 execs. It was the place I saw what happens when smart people actually care about the work.

One client ran a heavily regulated helicopter operation and was drowning in paper tomes. We spent days whiteboarding the real pain, doing UX work, and rapid prototyping. He got it approved and could finally spend more time flying than filing. That showed me what caring about the problem can unlock.

Another client needed perfect audio sync across phones at a party. The tech we came up with was incredible, but the founder ran the team like a blood sport. So I protected the builders, got people sharing instead of competing, repaired the relationship, and kept the product moving. I still use that instinct.

By year four at Boltmade, we had to stop being a family and grow up or sell. Google circled. Shopify fit. We joined in 2016, when Shopify had around a thousand people and Plus felt like a rebellion inside the giant.

My first year was "special projects": twelve problems in a row. I broke Shopify early-on—account invites overloaded databases and slipped past rate limits. Nobody yelled; they told me to fix it and learn. I started living on planes to Ottawa and Montreal. Relationships move hard problems faster than heroics.

I raised my hand for what everyone called "open-heart surgery": multi-currency. People said two years. We prototyped a version in a couple weeks; Montreal killed it on payments risk. I'd already been burned on money once before. I wasn't going to bluff here.

We built prototype after prototype, argued through arbitrage edge cases, and landed on MoneyBags: explicit currency context, backward-compatible casts, a migration path that didn't break the world. Launched in under a year.

Then came the long middle: a Plus dashboard that got axed, big swings at CMS, inventory, and CRM that failed. Same lesson every time—if you don't fix the core, big bets turn into theater.

That pushed us to the app shell: sidebar, top nav, what became Destinations, app pinning, Command-K, the early bones of notifications. Same people, same shape of problem. By the end I was burned out on the loop. I needed something new.

The reset was a new org; the role was sold as a year of hell. They weren't wrong, but I loved it. I inherited two messy teams—Metafields and Media—under leadership pressure, with no coherent story for either. The mandate was simple to say and brutal to run: finally ship a CMS.

After six months arguing about the minimum viable thing, I wrote the black-box rules for Metafields and the room snapped into focus. Metaobjects emerged; we raced and had them live in 6 months for Winter Editions. On Media we stood up a tiger squad and fixed the obvious rot first: WebP upload/download mismatches, bad quantization, CDN quality. Trust came back fast. When we layered generative-AI on top, the winning move was honesty—not "pick from six options," but "click a box, get a prompt."

That work spread. Metaobjects touched more of Shopify; Media got cleaner and more central. Then taxonomy arrived—a headline project with real weight. We rebuilt the PDP, shipped ML that balanced correctness with human taste, and took the harder path: ground it in Metaobjects so the core got stronger for everyone.

I made director after taxonomy. Near the end I helped shape The Catalogue, Shopify's first serious swing at a global view across stores. After eight years of bumping against shop-level walls, it felt like a real ending.

Early 2025 I was still riding high from Shopify, but GrowthLoop pulled on something deeper. I left to learn whether I could lead outside Shopify's machine.

The contrast was immediate: cash-constrained, too much to balance, and all running through one central meeting that had become its spine. Forty people, thin updates, everyone pretending things were fine while work slipped. Responsibility kept moving upward instead of landing anywhere real.

Fixing it was tougher than I expected. I killed that meeting and replaced it with projects and champions so ownership had somewhere real to land. Then came the slower work: building an internal operating system with AI, rewriting values, raising the talent bar, and pushing harder on debate, truth-seeking, and owning the hard calls.

The wild part was that, in an AI company, almost nobody was actually operating AI-first. Engineering moved first. Now the same shift is hitting the rest of the company. We're figuring out, in real time, what all of our jobs become when AI is the default. That's the story I'm in now.