I'm Kevin Korpi. I run Hold Fast Studio.
I design 0→1 AI products. The ones that touch the real world. Agents, sensors, hardware, vision, wearables. The stuff where the seams between disciplines actually decide whether the product ships or dies.
Been at this for 25 years. Microsoft, Google, Amazon Kindle, Android, Microsoft Research (where I was designing conversational AI in 2015, back when "AI" still mostly meant NLP), Moment, GoPro, Google Fonts, Adobe Frame.io. Three camera systems from scratch. A creator app I took from $0 to $1M ARR in a year.
Before all of that, a younger me was lead designer on Google AdWords, Google Analytics, and Google's first landing page experimentation platform. I don't show that work (it's from 2008), but the habits from those years (measure, iterate, ship, repeat) still run everything I do.
Running Hold Fast Studio since 2019. I've done a year at GoPro and six months at Adobe Frame.io along the way when a product pulled me in. Twenty years in and out of big companies taught me which type of designer I am: the kind who wants to be close to the work, close to the impact, and far from politics. So here we are.
What I actually do
Embedded design partnership, mostly. Founders bring me in early, when nothing is figured out yet, and when the thing they're building has to feel right, not just work. I work at 10,000 feet and 10 pixels in the same week. Strategy, product, UX, UI, motion, brand, systems, and when it's the right tool, code.
Two clients at a time. $15–25k/month, three-month minimum. Fixed-scope projects from $25k if that's cleaner for you.
Green flags: you're early, the problem is complex or physical or agentic, you want a partner who ships instead of making Figma decks, you care about craft, you can make decisions faster than every other week, and you want someone who'll push back. Clients tell me I'm human, blunt, and occasionally prickly. Guilty as charged.
Red flags: you need a career manager, a pixel-pusher, or someone to run the weekly design guild meeting. That's not me. (Type 1 designers are great. I'm just not one of them.)
Why the AI work
Honestly, my heart is in cameras and creator tools. GoPro, Moment, Contour. That's my love language. I still think about capture UX more than is probably healthy.
But cameras are a narrow market, AI companies keep reaching out, and the instincts trained on designing hardware that people pick up and hold turn out to be exactly the right instincts for AI products that actually feel good to use. Most "AI designers" showed up three years ago. I've been designing AI products since Microsoft Research in 2015, when it was more NLP than generative, and I've been designing things people touch for a whole lot longer than that.
So the AI work is the front door. The camera obsession is why I'm good at it.
Right now
Designing for a multi-agent computer-use platform that's beating industry benchmarks. Working with a growth and AEO product that plugs into CMS and analytics to generate revenue-moving pages on autopilot. Helping an AI medical platform simplify how new applications get onboarded into hospital ecosystems. A few VC incubators on the side. Occasional advisory to founders who hired design too junior, or can't yet afford a full-time designer.
Where I'm writing from
Bend, Oregon.
I grew up on skateboarding, snowboarding, and punk rock, and I haven't outgrown any of it. It's still in the DNA of how I work: learn the rules cold, break the right ones on purpose, commit to the line, own it when you eat shit. Ship it, don't pitch it.
I studied English, wanted to be a writer. First real job was as a copywriter in Tokyo, which is where I learned that writing and designing are the same craft pointed at different outputs. I still write a lot. I still snowboard more than is advisable. Cameras are always near me.
