Senior product designer with seven years at startups and scaleups, working end to end from research to shipped UI. I prototype in code with AI, and help others do the same.
Taking bookings in Stockholm, designed and built solo in a few weeks.
In short
An in-home wine, beer and whiskey tasting service. I designed and built it alone in a few weeks: brand, site, booking flow and back office. It looks like a marketplace, but I match every booking by hand, so there was no cold start to solve.
Lean on purpose
The obvious build is a marketplace with expert logins and self-serve booking. I skipped it. Mapping the flow between me, the customer and the expert first showed which steps to automate and which to keep manual, and manual matching turned out to be both faster to launch and better for the guest.
The booking flow across three parties. Every point where things could go quiet became either an automated email or a step I handle myself.
Design system first
I set the palette, type scale and component rules before building, moving between Figma and Claude through the Figma MCP. Slow at the time, but every page after that went up fast.
Colour, type, components and spacing, locked before the first page was built.The wordmark.
What the customer gets
Homepage, category pages, gift cards and guides, in Swedish and English. Swedish guests buy their own bottles, so a booking ends with a shopping list built by the expert.
The homepage.Pricing updates live as you add guests, so there is no surprise at checkout.Pricing in the open, before you book.Gift cards, redeemable in part.One of the editorial guides, which double as SEO.The shopping list. Each item carries its Systembolaget article number, so it links straight to the bottle.
The half I run
A separate back office moves each booking from new to matched to paid, with Stripe payouts and gift cards. Static site, Next.js back office on Supabase, Stripe, Resend. Soft-launched in Stockholm in spring 2026.
A booking from matching to payout. I designed my side and the customer's side together.Every transactional email, in one place.
Date
Apr 2026
Client
Waitwhile (internal tool)
Outcome
Designers now ship prototypes on real code without an engineer.
In short
Waitwhile designers prototyped in Figma, wiring every interaction by hand. I built a platform that clones the production code into a working prototype, so a designer can change the real product and publish a link with no engineer involved.
Real code, not a copy
The product is two codebases: an Angular staff app and React customer pages. One prototype can include both, so a flow that crosses them stays intact. The harder half was making an agent reliable in a codebase it has never seen, so I wrote a skill that tells Claude how our code works: the design system, the patterns, the business logic.
The gallery
Prototypes live in a retro Windows 9x desktop, and the metaphor does real work. It keeps them obviously separate from the product, and it is the one place every live prototype can be found.
Each prototype opens in its own window.Archiving is dragging it to the Recycle Bin.Pick the repositories and the slice of product; the gallery writes the prompt that sets everything up.New designers set themselves up from the built-in guide.
From prompt to live link
Designers already worked in Claude, so the loop ends there: say publish, and it goes live on your own dev URL. Every prototype gets a shareable link and a device switcher, which is what made customer testing on real code possible.
Starting a prototype from inside Claude.Published to a personal dev URL, no hosting or deploy step.The device switcher, mid-test.A live user test with an enterprise customer.The prototype our CEO and CTO presented to Allegion, the company that acquired Waitwhile.Click any element to pin a comment to it, so review happens in context.
Two decisions
Prototype on production code, because the interactions people argue about only feel right when the components and logic are real. And keep the hard parts hidden behind a prompt and a skill: the wins here were infrastructure, not button placement.
Date
Jan 2025 - Mar 2025
Client
Amici Event
Outcome
A beta the founders can learn from, instead of a year-long build.
In short
Amici had an events business and no product. I decided what to build first, convinced the founders, and designed it.
The bet
The founders wanted a full operational tool for event firms. I argued for one simple thing first: a single booking from a simple site. Smaller, faster, and we would learn more from it. I kept a login in it so we could add business customers later without rebuilding, and left AI out of the first release, shaping the data model so we could add it once there was traffic to build on.
The end user journey, mapped with the team before any screens.The internal process behind it: what the first release had to support, and what it could skip.Low fidelity first, so we argued about structure instead of pixels.
What we shipped
The full booking flow, from landing page to offer, designed with our developer. Every booking is handled by hand behind the scenes, so the team needed tools of their own.
The landing page.The request form.Venue selection.Waiting states matter when a human is doing the matching.The offer, once it is ready.The account, kept in for business customers later.Admin: every request and its status.Building an offer, chasing chefs and partners.Messaging, so the team works in one place.
Built to learn
It is a beta, so success is knowing what to build next: whether people want recurring bookings, whether they expect us to handle transport, which customers lean in hardest. A small design system keeps the flows consistent while we find out.
The design system, sized for a beta rather than a platform.A one-pager, so sales could pitch the service while the product found its shape.
Role
One designer, two non-technical founders, one developer. I made the product calls and worked with the developer on where to cut corners and where not to.
Date
Jan 2023 - Sep 2023
Client
Waitwhile
Outcome
Settings rebuilt in-house, with support tickets down almost 20%.
In short
Waitwhile was rebuilding settings in-house, which was my opening to fix the structure: long pages you scrolled forever, and a support queue that showed it. I tested three structures with users and shipped the winner. Tickets dropped almost 20%.
Three structures, tested
I started from usage data, support tickets and session recordings, then built three versions: the long pages we already had, a side menu, and a landing page you drill into. The drill-down won clearly. People look for a setting the same way every time: find the area, go in, change the thing, leave.
Where people got stuck, from usage data, tickets and recordings.The three structures, built far enough to test properly.Tested against the live product, not against each other.The drill-down won on every task.
Building it
I rebuilt the information architecture around workflows like staff availability and service management, on top of a component system, then tested it internally and with enterprise customers on live calls.
The full set of pages, wireframed so we could feel the structure end to end.Specs for every breakpoint.The landing page: find your area, go in.Editing hours without leaving the page.Table components, reused across settings.A settings page in the new structure.Advanced settings, out of the way but not hidden.
Outcome
Settings tickets fell by almost 20%. Time in availability settings went from 5m 20s to 4m 05s, about 23% faster. New modules now ship quickly on the component system. I was the sole designer, with a PM and three engineers, and drove the call to rebuild the structure rather than reskin it.
Experience
Waitwhile
Senior Product Designer
Feb 2026 – Present
Lead design end to end across research, UI, and prototyping, drive AI-centered design initiatives, and own the design system for a SaaS platform used by 4M+ people monthly.
Product Manager
Aug 2024 – Feb 2026
Set product vision and expanded the enterprise suite with App Clips, surveys, and a location directory. Introduced beta-state features and a framework for piloting experimental work with our largest accounts.
Product Designer
Jul 2022 – Aug 2024
Owned UX and UI for SMB and enterprise customers. Rebuilt the settings information architecture and a reusable component system, cutting settings-related support tickets by ~20%.
Amici
Design Lead & Board member
Jan 2025 – Present
Lead product design for an AI-powered food-tech platform connecting chefs with event professionals. Designed the end-to-end booking flow and admin tools, and structured the data model for future AI.
Hemkonst
Founder
Jan 2021 – Feb 2022
Founded a curated marketplace reselling original art from emerging Swedish artists. Designed it in Sketch and built the entire product in Webflow solo, reaching break-even within 10 months.
Påmind
Product Designer & Product Owner
Jan 2019 – Jun 2022
Led product and UX for a fintech loyalty platform serving banks, insurers, and telecoms, driving an 82% increase in Power of Attorney signups through UX improvements.
Daniel ReveszSenior Product Designer · Stockholm
