I set product direction with the founding team, hire and grow design teams, and ship AI-native products myself. Sixteen years, four companies, four design teams built or rebuilt.
Davey Miyabe (David Miyabe). Head of Design, Instil, Los Angeles. 16 years in product design. Open to Senior Director and VP of Design roles. Résumé (PDF): David_Miyabe_Resume_2026.pdf. Plain-text résumé for applicant tracking systems (PDF): David_Miyabe_Resume_2026_ATS.pdf
01 / Instil · Head of Design · 2020 to present
Instil
Result: Founding executive: a $15M Series A, 40+ clients, and AI-assisted delivery cut from about 5 months to 3 weeks.
Building design from zero as a founding executive
Fifth employee. Hired the design team, helped set the company’s culture, and led its first AI features.
Founding executive team
Design team of 4, hired by me; onboarded every company hire
Company grew to just under 40 people and raised a $15M Series A
In short
Set the product vision with the founding team, hired the design team, led the first AI features that cut donor briefings from hours to seconds, and rebuilt design around agentic development as the company shrank.
Before
We started with a pitch deck from the founders. The founding team (a board member and co-founder, the CEO and co-founder, the CTO, the head of product, the COO and me as head of design) researched the nonprofit landscape to build a product vision and a roadmap from nothing.
The platform
Product owned the roadmap and vision. I owned information architecture, design systems, branding and interaction patterns. Engineering owned infrastructure and security. The lines were gray. The CTO was my boss and not the final decision-maker, but he had a strong voice. When the head of product and I wanted to go a different direction, we built the case from our user interviews and a product-landscape analysis instead of opinion.
Most disagreements came down to how product direction affected data structure. Because we were serverless, flows had to be designed around that architecture’s limits, so design and engineering worked through the trade-offs together.
Image: Who owned what across product, design and engineering. (Diagram of product, design and engineering ownership at Instil)
Image: Instil Platform: the major-donor pipeline, and a donor profile on mobile. Information architecture, design system and interaction patterns were mine. (Instil Platform on a laptop showing a major-donor pipeline board, and on a phone showing a donor profile)
The hard call
The hardest call was research versus speed. I wanted more formal user research. The CTO argued that building something thoughtful and fixing it later was a better use of limited time and money. We kept interviewing, just less formally, and it cost us deeper insight into how best to support fundraising teams.
First AI features
I led our first AI features from concept through launch, marketing and sales. Donor briefings pull from the CRM with one click and take about 30 seconds, where they used to take hours.
Voice logging lets a major gift officer record meeting notes with a phone call on the walk back to the car: about 3 minutes instead of 15 to 30. The same assistant works by text.
Image: Before the meeting, a one-click briefing. After it, the assistant asks who you met and logs the notes. (A generated constituent briefing for a donor, with giving pattern analysis, engagement insights and strategic recommendations, beside a phone where the Instil Assistant logs a meeting by text)
Image: Hours to 30 seconds; 15–30 minutes to about 3. (Diagram: donor briefings and meeting notes before and after the AI features)
Design for agentic work
In 2026, slow sales shrank the company to three people and we moved to agentic development, and design’s job changed. I now work with product and engineering on an end-to-end AI product development system, and I build the design system and design ops that keep our design language and UX practices intact inside it. It cut end-to-end delivery from about 5 months to 3 weeks.
When code is generated, the system has to carry the judgment a designer used to apply screen by screen. So the rules get written down: one shape per job, colour only for status, contrast guaranteed by default.
Image: A design-system proposal I wrote for agentic development: one shape per job, so generated UI stays consistent. (Design-system proposal titled Pill grammar: one shape per job, comparing production chips with proposed status, tag, signal, link, filter, choice and button styles)
Result
Donor briefings went from hours to about 30 seconds, and meeting notes from 15–30 minutes to about 3. I hired a design team of four, onboarded every hire across the company, and set hiring and culture practices with the executives. The design language now holds up through agentic development.
Looking back
I would build a formal research practice sooner, using sales and customer success to get in front of nonprofit fundraisers earlier, and agree up front how to balance their goals against research needs.
Founding executive · Influencing without authority · Launching AI products · Rebuilding design for agentic work
Result: A $500K widget project became a $2M product, with same-dealer sales up 22.75%.
From a buggy widget to an end-to-end shopping experience
A $500K widget project that became a $2M engagement and Toyota’s blueprint for working with vendors.
3 teams: SmartPath, Toyota Owners, Toyota Financial Services
14 designers, plus a dotted line to about 20 UX designers
Headcount across the teams set each contract’s budget
In short
Turned a $500K widget fix into a $2M end-to-end shopping product by aligning every Toyota department behind one plan, then made the way we worked Toyota’s blueprint for vendors.
Before
SmartPath began as Digital Garage, a widget on toyota.com that a vendor had built poorly. Saving worked in some places and not others, the widget didn’t always appear, and the UX didn’t match how shoppers think. Our brief: fold it into toyota.com’s interaction patterns, add saved deals, and bring it into the design system.
Making the case
Early work showed the client wanted more than a cleaner widget. They wanted SmartPath woven through the site and usable in the dealership. They resisted at first: it expanded the budget considerably and needed sign-off from many departments. My account lead and I got an estimate from the engineering team and presented it to the client’s boss. It was a smaller lift than they feared. Then we brought the remaining departments on board and built one plan that fit the client’s vision.
Image: Stakeholder workshop and sketching session: getting every department into one room. (Stakeholder workshop and sketching session)
A new way of working
We brought human-centered design and an agile way of working to the client and engineering teams. It became Toyota’s blueprint for working with vendors, and my account managers and I taught other team leads how to run it.
Image: Ideation: sticky-note walls, printed pages and whiteboard flows for saving, comparing and buying. (Collage of a sticky-note wall, a printed vehicle page and whiteboard sketches of saved vehicles, compare and KPIs from SmartPath ideation)
Image: Style tiles: four directions for SmartPath’s tone, from Energetic to Natural & Human. (Four SmartPath style tiles: Energetic, Energetic and Human, Natural and Human, and Natural and Human dark)
One room, one plan
SmartPath touched toyota.com, the dealership experience and Owners, so the client was juggling many stakeholders. With my account lead, I got them in one room to set a shared vision, then held regular check-ins. Approvals got much easier.
Image: How alignment worked: separate sign-offs, then one room, then one plan. (Diagram: from separate sign-offs, to one room, to one plan)
Result
$500K grew to $2M over a year and a half, and Toyota could offer an end-to-end shopping experience: the first of its kind to let shoppers buy a car online.
Launched with two pilot dealerships, expanded to about ten, then rolled out in waves. When COVID hit, it let Toyota keep supporting shoppers from a distance.
Same dealers, before and after: sales up 22.75%, gross profit up 42.55%. About 60% of shoppers took the first offer.
I creative-directed the digital implementation of Toyota’s Start Your Impossible Olympics campaign, led by a designer on my team. It won a Cannes Gold Lion and the 2019 Grand Prix at the Clios.
Image: SmartPath on toyota.com, and the dealer tablet that picks up where the shopper left off. (SmartPath on toyota.com and on a dealer tablet)
Image: What it produced: budget, same-dealer results and rollout. (Diagram of SmartPath budget growth, same-dealer results and rollout)
Image: J.D. Power: “Toyota’s new SmartPath functionality removes friction from the car buying process.” (J.D. Power quote and SmartPath results: sales up 22.75 percent, gross profit up 42.55 percent, with verified user quotes)
Looking back
I would have aligned earlier with my own leadership on my plan to work across departments, so we could scale SmartPath’s approach together.
Multi-team leadership · Changing how a client works · Stakeholder alignment · Business results
03 / Second Spectrum · Creative Director / UX Design Manager · 2014–2017
Second Spectrum
Result: First designer to a team of 9, with tools adopted league-wide across the NBA, ESPN, Turner and Sky Sports.
Designing what a sports broadcast shows when computers can see the game
First designer at a computer-vision startup. Three years later: a design team of nine, and products used weekly by ESPN, Turner Sports, Sky Sports, and leagues around the globe.
First designer, grown to 9 full-time and up to 15 with contractors
Motion, UX, video editing, illustration and 3D
Partners: NBA, MLS, Premier League, Bundesliga, ESPN, Turner, Sky Sports
In short
Hired a multidisciplinary design team from scratch, partnered with data science and the CEO, and chose to ship CourtVision on time rather than perfect every augmentation. It led to the NBA, ESPN and Turner.
Before
The company was the founders and a few young engineers. There was a computer-vision technology and no design function. Nobody had decided what the product actually was, or who it was for.
Building the team
I hired for the sports and entertainment industry, where customers beyond the teams expect a high level of polish. The first hire was a motion designer with UI experience. As the work widened, I hired designers who could speak across disciplines (editors, motion designers, UX designers, illustrators, 3D artists) so one team could support every product we shipped.
Image: The team, and who used the work. (Diagram of design team growth and the partners it served)
Getting the data right
We were inventing metrics with computer vision and machine learning, so I worked with data science and engineering to make sure every graphic represented the data correctly. I worked with the CEO to support fast-moving deals and shape the company’s direction, including videos, slides and prototypes for a TED talk.
Image: StoryBoard: testing visualizations that build on mental models fans already have. (Six shot-chart visualization explorations for StoryBoard)
Image: ClippersIQ: live analytics and a trivia game for fans in the arena. (ClippersIQ app screens with a fan leaderboard and player shot charts)
The CourtVision call
The defining call came with CourtVision, the Clippers’ interactive augmented streaming platform. The CEO wanted a certain number of augmentation views, and the product manager needed us to finish so we could test and ship to the Clippers, media outlets and the NBA. I could spend months refining the look of each augmentation, or get the full set live by launch. I shipped. Some of the creative was weaker than I wanted, but it got the message across, and we improved it release by release.
Image: The CourtVision call: ship the full set on time, improve release by release. (Diagram: shipping CourtVision on time and what it led to)
Image: CourtVision: augmented graphics themes fans could switch between, live. (Four CourtVision broadcast frames with different augmented graphics themes)
Result
CourtVision ran in every Clippers home game not on national television, out of 41 a season, and grew into a broadcast tool for ESPN and Turner.
The analytics tool served every NBA team and MLS club, plus Premier League and Bundesliga clubs. Sky Sports used the team tool. It has since expanded to J.League, La Liga and more.
ESPN and Turner used StoryBoard weekly. The Clippers signed on for ClippersIQ and DataFX after seeing the analytics platform and our prototypes of the new metrics.
The suite was adopted league-wide across the NBA, ESPN, Turner and Sky Sports, the company raised a $20M Series B, and the work helped it win multiple Sports Emmys.
Image: StoryBoard 1.0 on ESPN’s NBA Countdown. (StoryBoard version 1 on air comparing two players’ shot charts)
Image: Concepts we pitched to take it to other sports: Monday Night Football and MLS. (Concept mockups of StoryBoard-style graphics for Monday Night Football and an MLS shot chart)
Image: Player augmentations on a Premier League broadcast. (Player augmentations on a Premier League broadcast)
Looking back
Engineering ran the company and I was the only design-focused leader, so speaking up for the user fell to me. I would do it earlier and louder.
Building teams · Cross-functional partnership · Executive influence · Shipping under pressure
04 / Liferay · Global Design Manager · 2010–2014
Liferay
Result: Built the UX practice: design team from 2 to 7 as the company grew from about 50 to 700 people.
Turning a marketing design team into a company’s UX practice
Joined at about 50 employees. The company grew to roughly 700 people, and design grew with it.
Design team from 2 to 7 at HQ, plus a part-time localization designer in Germany
Marketing support for eight offices worldwide
An annual conference that grew to about 2,000 guests
In short
Used a UX certificate and a customer-centered case to turn a brand team under the CMO into the company’s UX practice, with a design system adopted by every global office.
Before
I joined as a marketing designer. Design was a brand team reporting to the CMO, with no product design practice, and decisions came down to the CMO’s taste.
What I did
Owning the marketing website meant owning part of the portal, which exposed the need for real UX. I went back to school for a certificate in Usability and User Experience to get the foundation and the language. I worked with the CMO on the website architecture and made the case for a customer-centered approach over a marketing-led one. Walking through what that would look like built buy-in for UX, and I then worked with the UI team to extend it into the product.
I grew one designer into the manager who ran the team day to day and set creative direction. We supported marketing in eight offices worldwide, where the hard part was holding one brand voice across markets at very different stages of maturity.
Result
A design system adopted by every global office. An annual symposium and devcon that grew to about 2,000 guests. UX recognized well enough that the company created a formal design department.
Image: Liferay, 2010 to 2014. (Diagram of Liferay growth from 2010 to 2014)
Looking back
Sitting under marketing, I made its goals our main objective. That kept us aligned, but it limited the team’s reach. I would have looked sooner for places where design could help other departments, especially the product.
Building a practice from zero · Growing a leader · Global brand consistency
05 / Design Lab · Personal projects · 2026
Design Lab
Result: Two AI-native products, designed and built solo: HoopsAI and Nudge.
Shipping AI-native products myself
Where I design, prompt, code and launch my own products, to learn firsthand what AI-native product design takes.
HoopsAI: Shooting Coach, a computer-vision basketball training app for iPhone
Nudge, visual editing for websites founders have already built with AI (nudge-it.app)
Designed and built solo, with AI in the loop from research to code
In short
I lead design for agentic development at Instil, so I build with the same tools myself. The Design Lab is where I test ideas end to end, without a team to hand off to.
HoopsAI
A computer-vision shooting coach that turns a phone into the kind of tracking I designed for at Second Spectrum. I designed the product and its marketing site, and built it with AI in the loop. Status: [STATUS].
Nudge
Founders who build with AI can ship a site fast but struggle to tweak it after. Nudge lets them edit an already-built site visually, without re-prompting. The waitlist is open at nudge-it.app.
What it changes
Building it myself tells me what “done” costs with AI in the loop, where the tools break, and what a design team needs to keep quality high when the code is generated.
AI-native product design · Computer vision · Building in public · Hands-on craft
Intelligence, taste, culture fit and a willingness to learn. For senior hires, intelligence means framing a problem space and saying how success will be measured. Taste I read from a portfolio: high-quality work across different sectors and audiences.
02 How we critique
Early in a project we review often, to converge on a direction quickly. As the work fills in, reviews shift to the system and its architecture: zooming in on craft, then out to systems thinking.
03 Growing people
A designer I led had moved from marketing into UX. They communicated well but their visual skills lagged. We built those skills together, and they grew into leading the team and working with the client and engineering. Today they are a UX and service designer at a large consulting firm.
04 Prioritizing
With engineering and product, I find which projects block other work, which are high impact for low effort, and which have high leadership visibility. Then we stack-rank the work against our resources and each designer’s strengths.
“Always handled “fires” and “situations” with grace and logic … looked out not only for his team but for others as well.”
“Leadership style of placing people first, and knowing how to navigate company culture changes by forming genuine relationships across the organization.”
I teach and mentor outside my day job: adjunct professor at Cal State Long Beach since 2022, teaching user-centered and mobile design to up to 24 students a semester, and volunteer mentor with LA-Tech.org.