Stephen Sherwood — Product Designer

About

The longer version

Product designer with 3+ years of end-to-end experience — from blank Figma file to shipped product. Founding designer at a SaaS edtech startup, Design Methods TA at the UW iSchool, and currently building a solo product.

I'm at my best in environments where design decisions have real consequences and where someone needs to champion the user when it's inconvenient to do so.

Outside work: I bake. Inspired by The Great British Baking Show. Trail mix cookies are very good. Butterscotch pie is improving.

Currently: Building habit.all. Open to product design roles at stable companies with a real design culture.

I've been a founding designer, which means I know what it feels like to navigate the gap between "we need a design system" and "we have three engineers and a Tuesday deadline." I've done it with user research, stakeholder buy-in, and offshore teams — sometimes all three at once.

My background is Informatics (HCI) from UW, with a double in Comparative Philosophy — which turns out to be useful when you need to defend a design decision to a skeptical CEO and an equally skeptical engineering team simultaneously.

I TA'd Design Methods at the UW iSchool for two professors at the same time — reportedly the first undergrad in the department's history to do so. HR had to invent a new job code. The two professors had completely different pedagogies: one Figma-heavy, one writing-intensive. I learned to operate in both modes, which is probably why I write unusually thorough design documentation.

Outside work: I bake. Inspired by The Great British Baking Show. Still in the "follow the recipe faithfully" phase, but the trail mix cookies are genuinely excellent.

Currently: Building habit.all, an ambient data art product for smart home displays. Open to product design roles at companies with more than 10 people and a functioning design culture.

Founding designer at a SaaS edtech startup — which is a great way to learn exactly how much a blank Figma file can intimidate you before you stop letting it. Built the design system, shipped the LMS, wrangled offshore engineers, mentored four design interns. Roughly in that order of difficulty.

UW Informatics (HCI) + Comparative Philosophy. The philosophy part is actually just useful for arguing with engineers. Also TA'd Design Methods for two professors simultaneously — the iSchool had to create a new HR code for it, which I consider a minor institutional achievement.

Now I'm building habit.all solo — WebGL renderer, agentic AI architecture, HealthKit integration, and the slow realization that "solo founder" is just "founding designer" with even fewer people to blame things on. Enjoying it anyway.

I bake. The Great British Baking Show is aspirational television. I'm in the "follow the recipe" phase. The butterscotch pie is improving.

Currently: All-in on habit.all. Open to interesting work at companies that have survived long enough to have a design team. Done the founding thing — ready for something with slightly more structural integrity.

Work

Selected projects

Smashboard interface showing events calendar and progress tracking widgets

SaaS · LMS Founding Designer 2021–2024

Smash.technology — Learning Management System

Shipped a SaaS LMS MVP in 3 months. NPS improved from -27 to +46. Saved $200K via product strategy pivot.

View case study

Production was stuck. The offshore team was building the wrong thing, the CEO was running out of runway, and I was the only designer.

I diagnosed the failure mode, pivoted the product strategy to an open-source foundation, rebuilt the design system from scratch, and shipped a working LMS in 3 months — after two previous contractor attempts had failed over 8 months.

Outcome: NPS -27 → +46. $200K saved. First revenues generated. 4 design interns onboarded to the system within their first week.

Read the full case study

The company had burned through two contractor teams over 8 months before I proposed starting over with an open-source LMS as the foundation. The CEO said yes partly because we had no other options.

Three months of design system work, offshore negotiation, user research in Thai restaurants, and one very opinionated decision to kill the Smash Score feature.

Outcome: NPS -27 → +46. $200K saved. Shipped. The company didn't find product-market fit, but the product worked — and I have the design system receipts to prove it.

[Designer take]

The design system was the real unlock — not because it made things prettier, but because it let interns contribute without breaking things, which meant I could stop being a bottleneck.

Read the full case study

UX Research Teaching 2018–2022

UW iSchool — Design Methods Teaching Assistant

TA'd Design Methods for two UW iSchool professors simultaneously — the first undergrad in the department's history to do so.

View case study

Most TAs work for one professor. I worked for two at once — pedagogically different enough that HR had to invent a new job code to make it possible.

One class was Figma and Miro-heavy. The other was writing-intensive. Code-switching between two completely different modes of design thinking turned out to be excellent preparation for startup life.

Read the full case study

Teaching design methods is a reliable way to discover which parts of your process you actually understand versus which parts you do on instinct and call judgment.

Two professors, two completely different pedagogies. One wanted Figma deliverables. The other wanted thousands of words of written analysis per week. I wrote a lot of words.

[Designer take]

Grading other people's design work made me a much harsher critic of my own. Recommended.

Read the full case study

Showcase

habit.all

An ambient data art app for smart home displays. Connects personal data streams — health, weather, calendar — to generative visual artwork rendered in real time.

See the project

I got tired of dashboards that looked like spreadsheets. So I made one that looks like a painting that changes when your heart rate does.

habit.all connects to personal data streams — Apple HealthKit, weather, calendar density — and generates real-time visual mutations in digital artwork displayed on smart home screens like Samsung Frame. It's a solo founder project: product strategy, technical architecture, and design all mine.

The argument it makes: data has an aesthetic life. Utility and beauty are not opposites.

See the full project

I got tired of dashboards that looked like spreadsheets. So I made one that looks like a painting that argues with you.

The architecture uses a multi-modal agentic harness: an orchestrator routes tasks between specialized sub-agents — data ingestion, art generation, parameter mapping, rendering, billing — each with typed input/output contracts. Personal data stays on-device. Only prompts and billing events touch the cloud. Renderer is WebGL, cast via AirPlay/Chromecast.

Users generate custom artworks via text prompt, billed on token-usage-plus-margin alongside a freemium tier. Currently at prototype stage, which means I'm simultaneously debugging shader code and writing GTM strategy. This is fine.

[Designer take]

Building this alone is the best design education I've had since the founding designer role. Every decision is mine, which means every mistake is mine too. Highly recommended.

Full technical breakdown

Process

How I actually work

Understand the real problem

Not the stated problem — the real one. I do user research and take it seriously, but I'm also watching for the gap between what people say and what they do. That gap is usually where the interesting work lives.

Map constraints before generating ideas

Engineering bandwidth, business model, timeline, and the three things leadership will never approve — I want to know all of this before I start generating ideas. Constraints are not obstacles to design. They're the brief.

Design the system, not just the screen

A beautiful screen in a broken system is not good design. I build component libraries and design systems because they're the difference between a product that scales and one that accumulates debt.

Ship, measure, repeat

I'd rather ship something good on Tuesday and iterate than ship something perfect in three months. Perfection is usually a way of avoiding feedback.

Talk to people until I understand the actual problem

This sounds obvious. It is not obvious in practice. The actual problem is usually not in the brief. I conduct user research, synthesize it, and defend the findings to people who would prefer I confirm their assumptions.

[Designer take]

We called our synthesis sessions "rapid synthesis workshops." They were a whiteboard, two hours, and a lot of arguing. They worked every time.

Map the constraints before falling in love with solutions

At Smash, the constraint was offshore engineers on a waterfall contract with a Tuesday deadline. The design had to work within their existing infrastructure or it wasn't getting built. I'd rather know that on day one than day thirty.

Build the system, not just the thing

Design systems are not glamorous. They are, however, how you make it possible for other people to contribute without making things worse. I built one at Smash so interns could contribute meaningfully in week one instead of month three.

[Designer take]

Atomic design is a great idea that most teams implement in a way that makes everyone miserable. We took the spirit of it and threw out the parts that required three meetings to explain.

Ship it and actually look at what happens

I use data analytics. I also use the informal version: watching someone use the product for 5 minutes without coaching them. One is for stakeholders. The other is for me.

Let's connect Let's talk Say hello

Open to product design roles at stable companies with a real design culture. Best way to reach me is email.

If you're building something that deserves better design and you want someone who'll push back when it matters — I'd like to hear about it.

If you have opinions about design systems, generative art, agentic architecture, or the right way to make butterscotch pie — or if you're building something interesting — let's talk.

stephensherwood@outlook.com LinkedIn /now