End-to-end
Research through shipped UI. I'm comfortable owning a problem from the first messy conversation through to the spec an engineer builds from, and I prefer it that way.
I design B2B SaaS products end to end. Research, flows, interface, and the system holding it all together. Right now that's at Fullstory, a behavioural data platform, which means I spend my working life watching how people use software instead of how we hoped they would. It has ruined my ability to argue from opinion, which I consider a win.
01Selected work
All three shipped inside companies, so the full write-ups sit behind a password. The NDAs are real. Ask me and I'll send it over, or I'll walk you through the files properly, which is the better conversation anyway.
Knowledge Hub
Sellers waited months to get value out of the product. We cut that to weeks.
Lead Product Designer · 2024–25
Read the case study →
Call Insights Report
Nobody would install the integration without seeing value. The value needed the integration.
Main Product Designer · 2022–23
Read the case study →
Marketplace
One surface, two audiences: marketers installing apps and developers building them.
Senior Product Designer · 2021–22
Read the case study →
Public work: Emoot, a Vancouver fintech I took from brand through landing page to product UI, and a brand, packaging and print archive from the first half of my career.
02How I work
Research through shipped UI. I'm comfortable owning a problem from the first messy conversation through to the spec an engineer builds from, and I prefer it that way.
I narrow ambiguity in writing. A one-page problem statement with the constraints named has killed more bad projects for me than any critique session.
Lightweight moderated sessions plus behavioural data. Five users and a funnel beat a strong opinion, including mine.
Early and in their language. I'd rather lose a nice interaction in week one than find out in week four that it costs a sprint. Handoff is annotated and specced.
Components are a contract, not a library. I'm building an agent-ready design system, documented well enough that a new designer or an AI agent can use it correctly without asking me first.
I build Claude skills that automate the repeatable stretch between a PRD and a shipped screen, so senior time goes back to the product problem instead of handoff.
03Design ops & automation
Alongside product work I build Claude skills for my team. They go after the software-factory side of design: the repeatable steps between a PRD and a shipped screen that eat a designer's week and produce nothing anyone wants to look at.
Design ops is where senior time leaks. Every hour I take out of spec-wrangling and handoff is an hour back on the actual product problem. And it compounds across the whole team, not just me.
A design loop that runs like a factory line where it should be repeatable, and stays human where judgement actually matters. Part of that is an agent-ready design system: documented thoroughly enough that an agent can build from it correctly, not just a human.
Read: PRD to Pixels, the 3 Claude skills that fixed my design loop ↗
04Experience
Senior Product Designer at Fullstory, Klue and Unbounce. Three B2B SaaS products in a row, each with real users and real data behind it.
The through-line: I learned craft and restraint in print, then learned in SaaS that a beautiful screen nobody completes is a failed screen.
05About
I started out as a graphic designer, moved into web, and have been in product design since. Three countries' worth of working habits came with me, and they show up as a bias toward plain language and a low tolerance for decoration that isn't doing a job.
Day to day I work mostly in Claude and Cursor now. Figma and Illustrator are still here, but they've slid to second tier and I'm actively working out how much further I can push that. What holds my attention now is less which tool draws the rectangle, more how much of the work around it I can hand off. I work in English, Portuguese and Spanish.
Outside work I paint watercolour, spend more time than is reasonable looking for a good brewery, and keep a steady diet of pop art and contemporary work. Keith Haring above all. Bold line, flat colour, nothing hiding behind polish. It's close to what I want from an interface.
06Before you ask
Answered straight, including the ones that don't flatter me.
I'm happy at Fullstory, and I'm not running a search. But I keep the door open. If you're building something where the hard problem is making dense information usable, I'll always read the note and take the conversation. The worst case is that we compare notes and both learn something.
B2B SaaS, almost entirely. Competitive intelligence at Klue, a landing-page platform at Unbounce, ordering software at Craver Solutions, behavioural data at Fullstory. I'm happiest in the dense parts: reporting, information architecture, onboarding, the screens where someone has to make a real decision from what they're looking at. If the hardest problem on your roadmap is making something prettier, I'll be bored by month three and so will you.
Teams who want a designer to receive decisions rather than help make them. I'll ask what a request is for, and if the answer is thin I'll say so early, in writing, before anyone has built it. Also: if there's no appetite for looking at how people actually use the product, I'm the wrong fit. I'd be arguing from data nobody wants to read.
Tell me the business problem, not the solution you already sketched. Give me access to real user behaviour and a sharp engineer to argue with. Then be specific in feedback. "This feels off" costs us a week. "This buries the thing sellers need in the first ten seconds" costs us an afternoon.
Finding the thing you can give away before anyone has committed anything. Both Klue projects came down to that: value first, plumbing second. It sounds like a growth tactic. It's really just refusing to ask people for effort before you've earned it.
Knowing when a system is documented enough. I'll keep refining a pattern library past the point of return because I find the documenting satisfying. I have to set a stopping rule and hold myself to it. I'm also deliberately shifting my practice onto AI tooling, which means I'm regularly the least expert person in a room I used to be comfortable in.
07Get in touch
I'm not running a job search, but I like talking to people who build things. If you want to argue about onboarding, compare notes on design ops, or ask how one of these projects actually went, write to me. I reply within two business days.
Vancouver, BC · Pacific Time