— Eddy Merckx

— Kurt Vonnegut

These are the projects that make me want to stay a designer who codes rather than pick one lane: understanding a problem well enough to prototype it in the browser changes the quality of the questions you ask.

If you're wrestling with something similarly gnarly, that's usually the fastest way to get my attention.

There's an obvious parallel to design work — you plan as much as you can, then adapt constantly once you're actually moving. The best routes, like the best products, rarely look like the ones you sketched out the night before.

Favorite trips so far have hugged the North Sea coast and wound through the river deltas of the Netherlands. Longer routes across Europe are next on the list.

Along the way I've lived and worked across Sweden, China, the Netherlands, and a handful of places in between. Each move reshuffled how I think about problems — less about finding the "right" answer, more about understanding the constraints well enough that the right answer becomes obvious.

Outside of work, I'm usually planning the next trip, fixing something that doesn't need fixing, or reading about a subject I have no business studying. This site is where all of that spills over.

Taken somewhere between a design sprint and a bike ride. Click through for the full-size version.

This site is my personal log: notes, case studies, and small visual experiments.

I led product design at Alphadoc, a DevEx tooling startup, from proof of concept through its first public release within a year.

The core of the product is a sequence-diagram builder: drag a request between a client, a server, and an API, then attach a real endpoint and method to that step. Diagrams could also be generated straight from an uploaded API spec, turning a step that used to take an afternoon of manual diagramming into a few clicks.

Alongside the diagrams, a block-based editor let teams write tutorials the way a CMS lets you write a blog post: text blocks, code blocks, callouts, and a dedicated API-reference block that pulls a live endpoint into the page instead of describing it by hand.

Research, strategy, interface, and design system come from the same pair of hands. Small teams rarely need a design department. They need one designer who can hold the whole thing.

Selected clients

• Humane Lighting, LumenRadio, 2019. Consolidating AirGlow's light commissioning into one workflow, from planning to install to daily use.
• Carbon Counting, CarbonCloud, 2018. A food climate-impact tool, from kitchen input to guest-facing menus.

Lightspeed sells Point-of-Sale and eCommerce software to small and mid-sized retailers. I worked across five product teams on iOS, Android, and web. The biggest piece of it was a design overhaul of the native Point-of-Sale, a product people knew about and didn't use.

Years of acquisitions had left the portfolio fragmented. A retailer could run into three different Lightspeed interfaces doing the same job: a POS in the browser, a tablet app, and a phone app. Each had its own layout and its own idea of how a sale works. The browser one was what people used, despite limited functionality, slow transactions, and the cost of running a till on a full computer.

The brief was to consolidate those workflows, build a foundation solid enough to bring new retail verticals onto Lightspeed, and lift adoption of the native apps. So the question was what was holding adoption back.

Answering it took user interviews, collaborative workshops, workflow analysis, and a pass through the usage data. The usage data changed the shape of the problem more than the interviews did. Ringing up sales accounted for 93% of everything staff did in the app, average Time to Transaction sat at about 125 seconds, and most sales were two items or fewer.

Nearly every real sale was two items, while the interface was built for the rare large basket. Staff were spending over an hour and a half a day inside the POS, so seconds of Time to Transaction compounded into something they felt. The interviews added two things the data couldn't show. Staff had to answer a customer's question and show a sales breakdown on the spot, and value add-ons were worth real money to a small retailer, but nobody remembered to offer them.

Breaking a sale into those five steps and cross-checking each against what staff were trying to do produced four design principles. They did most of the deciding from there on.

• Follow the walk. Be guided by the user's workflow instead of imposing one that doesn't match it.
• Freedom to choose. Don't jump to conclusions. Leave flexibility where the workflow needs it.
• Transparency. Match what staff and their customers expect to be able to see.
• Hidden values. Surface the add-ons people forget, because the delighters did matter.

The layout followed from the principles. The old screen led with the cart and pushed products into a narrow column. The new one hands the primary area to the product grid and quick actions. The cart and checkout stay put on the right, and navigation gets its own row at the bottom instead of competing for space.

The revamped POS went into a private beta with more than 30 stores and saw significant adoption there. Time to Transaction came down, which was the win retailers cared about. Add-ons went up too, since putting one into a sale no longer cost anyone time.

Freelance work from 2019, through Stephan Ye Design. Old screens, and a handoff problem I still run into.

LumenRadio builds wireless connectivity for IoT lighting. Its flagship product, the AirGlow™ smart light controller, connects lighting fixtures wirelessly and runs them as scenes. I was tasked with designing how every stakeholder, from light designers to installers to the person flipping the switch, interacts with it.

AirGlow turns ceiling lights, ambient lights, and spotlights into scenes: sensors, triggers, and effects combined into a single lighting state a room can call up on demand. The interface had to make that complexity easy to set up and use, on a laptop for the person planning a project and a phone for the person installing it.

The project started with a strategy for flexible light commissioning, one that could adapt across different project needs and light openings. Stakeholder interviews surfaced where the existing flow broke down: unclear transitions between design and install, design intent lost along the way, setup steps that were too complex.

Consolidating the workflow onto one platform gave every stakeholder a consistent experience and cut installation and calibration time. It also made troubleshooting easier and improved handoff across the three-phase Light Commissioning process: planning, installation, and daily use.

I led product design at Verifone, a global payment service provider, across three workstreams over about a year: merchant onboarding, the payment gateway itself, and the company's first design system.

Onboarding started with a simple question, how a merchant takes payments, then walked them through picking a service tier and configuring it, the monthly price updating live as they added payment methods and picked a contract length.

The gateway underneath supported dozens of payment methods: cards, wallets, and regional schemes, from Visa and Mastercard to iDEAL, Alipay, and WeChat Pay, so a merchant's checkout matched however their customers actually wanted to pay.

Merchants tracked payments through a transactions dashboard: searchable by ID, filterable by date and organization, each row showing payment method and status. A CSV export handled reconciliation.

I also built Verifone's first design system, color and type scales, input and selector states, buttons, and message patterns for alerts and toasts, so product teams pulled from one shared source instead of rebuilding components per feature.

Freelance work from 2018, through Stephan Ye Design. The craft has moved on since. The problem hasn't.

CarbonCloud, led by food and climate researchers, helps producers and restaurants calculate the climate impact of what they serve. As a Product Designer, I spent 6 months turning customer research into a tool chefs could use daily and diners could trust.

Kitchens are unpredictable. Dishes change with what's in season, what sold out at lunch, what a supplier swapped in. Chefs needed to log climate impact as fast as they built the menu, not as a separate chore afterward.

The dish editor became mobile-first: add an ingredient, watch the CO2 estimate update, done. The same editor scaled up to a full week's planner on desktop, for the manager laying out the whole menu at once.

On the restaurant side, a dashboard turned that same data into something the business could act on, emissions, wasted portions, and revenue side by side.

Once a menu was set, it had to reach the people eating it. The web app translated each plan into printed menus and the guest-facing app, shareable via API, so a diner choosing between two dishes could see the climate difference before ordering.

Designed infotainment and mobile app experiences at Lynk&Co, working closely with engineering and product teams. Introduced Sketch and built the design department's first digital design system, improving consistency across platforms.

Project images are under NDA, but happy to walk through the process and outcomes in person.

Marina enthusiasts spend most of their time maintaining, storing, and servicing their boats, leaving little time for actual boating. That trade-off didn't sit well with a speedboat brand built around electric boating.

The goal: an integrated storage and charging facility for the boat itself, one that could work independently of existing marina infrastructure, self-charge, and keep the boat secure.

Development moved through several iterations, starting with a survey of marina facilities and existing boat-storage solutions. From there, a concept took shape: a canopy berth that docks, shelters, and recharges the boat in one motion.

The result was an integrated charging and storage berth for the Edorado 7.120, plus a standalone charging station for eco-resorts running fleets of electric speedboats.