Product and business thinking
I score the jobs before anything gets built. I argue funnels, retention loops and pricing gates in writing, then defend them. I arrive with an opinion about what the business needs, not a list of screens to draw.
I design products end to end and ship them in code, with AI agents as my build team: the decisions are mine, and every one traces back to evidence. Anything still unproven is marked [?] rather than smoothed over.
I score the jobs before anything gets built. I argue funnels, retention loops and pricing gates in writing, then defend them. I arrive with an opinion about what the business needs, not a list of screens to draw.
Entities, flows, sitemap, then a written spec for every page and every state. Nothing appears for the first time inside a wireframe. A prototype renders a finished structure; it does not invent one.
I build the thing. Design systems, prototypes and whole products, shipped with AI coding agents that I direct and review. Tendd went from a design file to a live product in five weeks. This page is one of them too.
Twelve stages, run in order, on every product. Each one leaves a written source of truth and a defect log behind it. Slower to start. Far faster to finish, and much harder to argue with.
Forgotten subscriptions cost the most to the people who never look. I designed a tracker they can bear to open, then built it and put it live.
Coaches buy supplements for twenty athletes by price list and phone call. I designed the store that is built around them.
An AI agent closes security alerts for forty clients at once. I designed the console that shows the analyst how far to trust it, client by client.
A legendary Counter-Strike skin gets a premiere, not a thumbnail. A film the reader plays by scrolling, with no cut the eye can see.
More work · design sprints, public end to end
Product managers lose the roadmap argument to the loudest voice in the room. The research and architecture of a tool that turns feedback into evidence they can defend, built on 19 customers' own words.
8 of 12 stages · visual layer in progress Case study
In a team of three, “anonymous” is a lie employees can count. The research and architecture of a wellbeing tool where privacy is the structure, not a setting.
9 of 12 stages · visual layer in progress Case study
Prediction markets are built for traders. Yonder is designed for people who follow the news and want a stake in it, with no wallet until the bet.
12 of 12 stages · complete Case study
Every portfolio says “I build in code”. This one lets you feel it in five seconds: a plush dragon whose head follows your pointer, who changes costume for each stage of how I work, and who answers when you poke him.
There is no rig. His head is a video and your cursor is the playhead.
A hole found while building a wireframe is not patched in the wireframe. It is fixed in the architecture, then rendered again. That single rule is why the visual stage is fast: by the time colour lands, there is nothing left to decide except how it should feel.
When I do not know something, I write it as [?] and carry it forward with the test that would close it. An unmarked guess is the most expensive object in a design file: it looks like a decision, it gets built like a decision, and it is only found once the product is live.
I have been designing digital products for more than ten years. Most of that time was spent inside one large B2C platform built on complex game mechanics, retention loops and reward systems: the kind of product where changing a single progression rule moves the behaviour of the entire user base, and where you learn that a screen is never the unit of work.
That taught me to design systems and to expect every decision to be argued in numbers. What I do now is the same discipline, run in the open and end to end: take a product from the first research question through to a working interface, and write down why at every step. The case studies here are not retrospectives written after the fact. They are the actual artifacts, with the defect logs and the open questions still in them.
I work fastest with teams that want an owner rather than a supplier: someone who will argue about the business model on Monday, restructure the information architecture on Wednesday, and have it running in a browser by Friday. Most of that speed now comes from working with AI agents: I make the calls and review what they write; they do the typing.
I am a team player who is comfortable owning the decision. The design happens inside a team: thinking shared early, critique taken straight, the craft bar raised with the people around me rather than above them. When a call has to be made, I make it and stand behind it. This far in, the work I still want is the design itself: the research, the architecture, the interface and the code. That is why the seat I am looking for is an individual contributor one.
If you are hiring for a senior product design role, or you have a product that needs to be taken apart and rebuilt, tell me what it is and what is going wrong. A rough description is enough. I will tell you whether I am the right person for it.