Derek Wood: Designer

→

People don't always know what to make of me. My path has gone in a lot of directions - breadth and depth, as you'll see below. I can write full-stack code, lead design at a startup, plan and build an app end to end - but what I want most right now is to be on a team doing something that matters.

→

CodePen embeds did that job well for years. But when we finally decided to build a robust internal replacement, we wanted it to tick all our boxes - including running server-side code (PHP, Python), which almost no browser editor does. We even added touches like CSS grid guides, and we keep improving it as we run across interesting situations. It's built as a standalone system that can be hooked up to any language runner.

→

IA conferences often turn into website-centric wireframing and UI talks - and conferences in general lately end up being all about "AI" - so I was set on keeping this one about information architecture itself. I defined three pillars for the theme and wrote the call for speakers around them. It was an interesting challenge to build a day around exploration and networking that was welcoming and thought-provoking for people of all experience levels.

Timelines, branding, graphics, the venue, the social plan, and so many shared logins I lost count. I sat with every speaker to help shape their talk for the theme and the crowd. But mostly it's people. A dozen volunteers giving their time for free, everyone with their own way of working, and my job was keeping them matched to what they're good at and moving the same direction.

→

The system breaks into four product scopes that tie together. Onboarding and agent-matching. Data upload plus education for getting the house and its documents ready. A course-like flow for being on the market - showings, offers, negotiation - paced so it feels complete and even a little fun instead of overwhelming. And a robust dashboard where agents and admins run everything behind the scenes. Four big flows, each with its own goals, and I built the whole thing. On a lean team, I treated every stakeholder as a user and a co-designer, and I made the tradeoff calls: what was worth building carefully in Figma, and what we could just build and refine right in the browser.

There's a lot of talk on LinkedIn about AI and 10x-ing this or that - but I used this project to explore everything Claude Code has to offer: project structure, an agentic design system, skills, agents, all of it. The real skill isn't using AI, it's judgment. From real experience, not hype, I know where LLMs and agents genuinely help - and where they become a trap that steals your human context, quietly builds tech debt, and erodes the communication a team runs on.

Like any large-scale project, launch waits on legal and marketing timing.

→

Being part of it was genuinely powerful. I spent days across the beautiful ArtCenter campuses, sitting among educators and provosts from the top design schools in the country - including CCA, where I did my undergrad. Two highlights: Safiya Noble's keynote (she wrote Algorithms of Oppression), and the Data Vandals, a data-viz art collective out of NYC. It was also eye-opening. Some of them were strangely out of touch about AI - one asked, in earnest, whether it means we could get students through a degree in two years instead of four. No. That's not how art school works, at all.

→

I'm not sure if it was COVID or something larger in society - but so many designers and developers were hiding out, not being human, not connecting. Part of the point was to get them to come out of their shell and meet someone who'd actually look at their work with them. People building robots. People with a bootcamp plus a CS degree plus a master's who still couldn't figure out how to do anything. Designers who felt down on themselves but turned out to be really talented once they learned how to talk about their work. Every conversation is different. It's not just about helping people, either - selfishly, I learn from every session. I've seen the inside of hundreds of people's ecommerce stores, visual design portfolios, and design system problems at work. That kind of exposure is its own education.

→

I came on as "the UI designer" - a title I'd never held on its own. I was honestly excited to see what it'd be like to work only in Figma, not as UX or dev. That lasted about a minute. They needed me on UX almost immediately, and from there I got pulled into every aspect of it.

The project had a little of everything: a unique stepped, data-driven onboarding flow, discovery for finding and joining your group, heavy API integration for video and meetings, and a hefty bit of data science for collecting the right data and matching people into pods. We also did a pretty deep R&D pass on a companion habit-tracking app.

They'd hired a brand designer to rework the identity. Interesting on paper - but the moment we put it on real screens, everything felt like Home Depot. I ended up redirecting the color system, rethinking the visual strategy, and rebuilding the brand guidelines so they held up in the product.

→

I ended up creating examples in every combination we could discover, documenting each one and making friends with blind testers along the way. We explored a bunch of ways to show the technology - scrolling storytelling, interactive demos - and I represented the work at the CSUN Assistive Technology Conference that year, where I met a lot of amazing people and got real feedback.

This kind of work goes deep - it's not a graphic you fix once and hand back. It's embedding with a company's internal team, shifting how they think about it, and working out how all of these pieces fit into their design and theming system at large.

→

Most font lists in style guides aren't helpful for developers - they're documenting something brittle. An h2 is a data-structure hierarchy, not a size. But since no one talks about it, you find things like <p class='h2'> to smooth over the incomplete thought. Type patterns are reusable text styles named by purpose instead of size - not "h2" or "extra-large" but "attention-voice" or "calm-voice." A shared language designers, developers, marketing, copywriters, and the boss can all use without needing to feel technical or creative.

Read the article

→

I'd worked with enough people from enough backgrounds to know: most of them couldn't work as a team, didn't understand cross-role goals, and had no shared vocabulary for how design and development connect. Coding bootcamps were charging $30k. I figured I could do something at least three times better for a third of the price. Lambda School raised $50 million and still failed its students. I started Perpetual Education with $50k.

I called the first program "Design for the Web" and had to frame it as a coding school, because nobody understands what product design is yet. It ran in cohorts - just two to three hours a day, six days a week, for six months, in contrast to the 10-plus-hour-a-day programs - and it got better results, because we made space for real learning instead of shoving React tutorials into your eye sockets.

I wrote the curriculum, designed the workshops, built the modules, wrote the glossary, made and edited the videos, created the animated graphics, handled marketing, and built a custom learning management system from scratch. The enrollment form was designed to teach people about the program while they filled it out, so by the end they knew whether it was for them. I met with every prospective student personally before accepting them.

We experimented with ISAs, payment plans, and different pricing models to keep quality education accessible. The platform grew over the years into a full system - milestone tracking, dashboards, and features we added as we naturally learned what we needed. The first cohort launched March 2021, with overlapping cohorts every three to four months for four years after.

→

Behind the scenes of these big movie companies is a ton of data. PXL had built custom content management systems and had all of that locked in. My job was to get DreamWorks in place and handle whatever else popped up - when the Bumblebee movie came out, we made a fun trivia game for it. There were always new projects and long-term contracts to balance.

Naturally I wanted to be on the fun stuff, so when I was put in charge of mentoring the others - a CS intern, a bootcamp grad, and someone who came over from marketing - I happily (and effectively) taught them how to do my job so I could focus on R&D. It was great to be back in a creative tech role after so much C-suite at the previous job. And it's what got me more excited about teaching than about climbing the developer ladder.

→

I designed self-tracking tools that let users monitor how they felt over time and draw their own conclusions. If people are feeling better and healthier, that's what matters.

I established the design surface for the company: their web presence, how they were being perceived, the component systems, the brand voice. I ran user testing through UserTesting.com, did competitive research, worked with the head of development, and hired the designers who eventually took over. Long before Charlie from Always Sunny, I had a wall-sized flow map of the entire business like I was tracking a killer.

→

The platform is essentially a web app that generates and maintains school websites from behind the scenes. Teachers put up work, students access projects, tons of data flows through it. My job was to look at the front-facing visual system and tell them if they were building their CSS right. Responsive design had arrived, Sass and Less were changing how people wrote stylesheets, and their team needed someone to evaluate whether their component architecture was keeping up.

Instead of getting into their giant Java system and scaring everyone, I used CodePen's ability to link pens within other pens and built the entire component architecture as a standalone, themable system. Same primitives, same components, same styles, but overridable per school. I met with every team member, mapped out who was writing what, and explained how it could all work.

I don't remember calling it a "design system" because I don't think people were using that term much in 2016. But that's what it was: a system of themable websites sharing the same primitives with per-school overrides. Whether or not I knew the word, I'd been doing this work for a long time.

→

The variety was the education. An interactive education game shown at the White House. Game-based insurance apps for AAA. Data dashboards with early ML for Fox Films. Brand microsites for Gap and Shell. Venue sites for Ticketfly. Every project was a different team, a different stack, a tight deadline.

Inevitably, these agencies would bring me on as a front-end developer, but I'd end up consulting across roles because responsive design was still new and most design teams didn't know how to design for all screen sizes. I'd help them understand how to make responsive work easy on the dev team, which usually led to them asking me to formalize their in-house process: build tools, how team members share information, how design and development actually collaborate. I was a big advocate for pair programming and cross-role working sessions. The bridge between design and engineering is where I've always ended up, no matter what the title says.

→

Up to this point, most of my work had been custom WordPress sites, brochure pages, content-driven projects. ShoutQ was different. It was a startup with investor money and a product that needed to work as an app: user accounts, matching influencers with brands, pricing, campaign management. I built the frontend prototype in Angular, handled responsive design and UI polish, and ran user testing.

The company was way ahead of its time. The influencer economy didn't exist yet the way it does now. They got overvalued, had investor issues, and eventually collapsed. But the experience of building a real app prototype, bridging design and engineering on a product team, was the turning point. After ShoutQ I knew I wasn't just a web developer who made websites. I was someone who could build products.

→

Great team. Learned a ton. The projects were full visual environments: Cam and Benny (an illustrated shelf-UI for animation creators), Murder Doll House (a dark theatrical world for a production company), Jamie Palumbo (a bold musician site). Each one was custom, responsive, and built to match a specific visual world.

I taught the team Git, introduced modern development practices, and started thinking systematically about components and reusable patterns before anyone was calling it "design systems." The agency eventually collapsed and offered me a partner role on the way down. I declined.

→

Fundamonium was the business name. Lucky to have had challenging early projects that forced me to dig deep. Time No Place (a record label site) was the "all was revealed" moment, where CMS, databases, and content relationships clicked into place as a design problem, not just a technical one. That project had artists, releases, shows, a blog, all connected. It was my first real WordPress custom theme and it changed how I thought about data.

Got really into responsive design right as Ethan Marcotte published the idea. Most designers were still building for desktop only. I was building for every screen from the start, and that early commitment to responsive thinking shaped everything that came after.