Dane O'Leary · Senior UX Designer + Web Developer

Timeless craft, shipped in modern code.

I'm Dane, a senior UX designer and web developer. I find out why something's broken, then design and build the fix, from healthcare signups to multi-brand design systems.

14 years in UX, design systems, and web · Healthcare, fintech, and nonprofit clients · Mentor with ADPList and AIGA NY

What I do

Design and engineering under one roof.

Most teams lose something between mockup and merge, but I operate on both sides of that line. Decisions made in Figma are reflected in production.

  • UX & product design

    Research, flows, information architecture, and testing that settles arguments. A location-first rebuild of a clinical-trial signup cut form abandonment from 65% to 37%.

  • Design systems

    Tokens, components, and the governance that keeps them honest. LINEAR runs ~266 tokens and ~75 components from one source of truth, rendered in both Figma and Webflow, with 360 brand skins from five theme axes.

  • Web development

    Webflow, WordPress, SvelteKit, and Three.js when the brief calls for it. I write the code, so the prototype isn't a promise someone else has to keep.

  • Brand & identity

    Strategy first, logo second. Vedara Ventures, a $100M+ holding company, got a complete brand in 14 days.

  • Accessibility & MX

    WCAG 2.2 AA is my floor, but I also structure sites AI agents can read with llms.txt, agents.json, and clean semantic markup.

My Projects

Check out my case studies in UX, design systems, and web development, including the processes and results behind each.

About Me

Why "Design Archaeologist"?

The most important thing I do happens before anything gets designed.

I dig through the analytics, the support tickets, and the 15,000-line stylesheet nobody wants to open, and I figure out what people actually did before deciding what they should do next.

I'm a history and archaeology nerd off the clock, too, so the name stuck. The method's the same either way: look at the evidence, ignore the legend.

I've been doing this for 14 years, across healthcare, fintech-adjacent startups, holding companies, and nonprofits. I design it, I build it, and I'll tell you honestly what didn't work on the way there.

Contact Me

Let's build the damn thing.

Need a website or mobile app that makes the complex simple? Are you hiring for a UX or systems-related role? Shoot me a 'text.'

Nonprofit Website

Kevin Popke Foundation: A Nonprofit Site Its Team Updates Without a Developer

PageSpeed 98 / 100 / 100 / 100

Problem
The old site buried Kevin's story, and every grant, event, or story update needed a developer the foundation couldn't keep on call.
Solution
I reordered the site as person, then purpose, then programs, and built a headless WordPress plugin that lists every heading and button as a plain editable field.
Impact
Launched August 24, 2026 on $2,000, scoring 98 desktop and 95 mobile for performance, with new grants added with no template edits or redeploys.
Client
Kevin Popke Foundation
Role
Lead designer + developer (solo, AI-assisted build)
Timeline
Concepts in week one; launched August 24
Open case study

Design System

LINEAR 2.0: Turning a 15,000-Line Stylesheet Into a Tokenized Design System

1,990 classes → ~400

Problem
LINEAR 1.0 was a 15,209-line stylesheet with 1,990 classes and variables, no dark mode, and no way to restyle a site without hand-editing classes.
Solution
I deleted the duplicates and rebuilt it as three-tier tokens, with matching class names in Figma and Webflow and accessibility built into each component.
Impact
About 400 classes instead of 1,990, project-wide light/dark and radius themes, and 86 of 86 color pairings passing WCAG 2.2 AA.
Client
VYNE Creative
Role
Lead designer & engineer (solo), VYNE Creative
Timeline
8 weeks
Open case study

Client Website

Location-First Design Cut Clinical Trial Abandonment From 65% to 37%

Form abandonment: 65% → 37%

Problem
Across four prior trial sites, 65% of people quit the screener, because they had to share health data before learning whether the trial was anywhere near them.
Solution
I put a searchable city directory ahead of the screener and saved facility matching for after qualification, so people could answer "Is this near me?" first.
Impact
Abandonment fell to 37%, verified participant matches rose 23%, and the client replaced its old template with this system on two more studies.
Client
SAD Clinical Trials
Role
Sole UX/UI designer (Lead Designer)
Timeline
6 weeks
Open case study

Client Website

312 Truck Website

Designed and built the first website for a commercial truck repair shop on Chicago's South Side. Daily estimate requests rose from 8 to 19, and revenue grew 13% within six months.

312 Truck Body Repair & Painting had no website, no analytics, and no lead capture, in a market where fleet managers pick vendors online and one job averages about $14,000. New ownership gave me six weeks to change that.

What I did

  • Designed desktop-first, because over 80% of B2B repair searches happened at work on desktop.
  • Swapped stock imagery for real shop photos after testers went looking for “real trucks” before trusting the business.
  • Built a quote form that explains each field and routes leads into HubSpot, cutting follow-up time from 15 minutes to 5.

Results

2.4× daily leads 59% form completion +13% revenue
8 → 19 estimate requests Goal was 35% Within six months

For the first time, a logistics company reviewed the shop’s website before a contract meeting.

Brand & Website

Brand and Website for a $100M+ Holding Company in 8 Weeks

Brand + site in 8 weeks on $7.5K

Problem
A holding company with 40+ businesses worth over $100M had no logo, no website, and no LinkedIn page, with investor conferences weeks away.
Solution
I designed the brand in two weeks, then built a WordPress site in six, with a CMS that adds a new portfolio company through a form.
Impact
Launched in 8 weeks on $7.5K with a 1.9-second load time, and the client now adds a company in 3–5 minutes with no training.
Client
Vedara Ventures
Role
Lead Designer
Timeline
8 weeks
Open case study

Mobile App

NAR Mobile App

Designed a mobile recovery companion for a nonprofit, built on encouragement instead of streaks and penalties. In a 35-person pilot with no incentives, 46% were still using it at day 120.

A nonprofit wanted a recovery app that motivated people instead of punishing a missed day. I led UX and visual design, working with people in recovery and two licensed counselors.

What I did

  • Built a dashboard that gives before it asks: a fact of the day, a poll, or a personal metric like “You’ve saved $362 since your quit date.”
  • Replaced numeric scores with plain status words (“Stable,” “Cautious,” “At Risk”) after testing showed a number reads as a grade.
  • Cut the daily check-in to two screens and put sponsor contact one tap away.

Results

35 volunteers 46% retention “Gentle”
Pilot, no incentives Still active at day 120 How testers described it

The client paused development before a wide release when subsidy funding fell through.

Concept App

Schedually Concept App

A self-initiated concept app for people juggling work, personal, and family calendars. A weekly template protects time up front, and all four testers set it up in under five minutes.

I was running three calendars, and client calls kept landing on doctor’s appointments because my work calendar said I was “available.” Four peers I interviewed had the same problem. They didn’t need better availability detection. They needed boundaries.

What I did

  • Flipped the model: define a weekly template of work, family, and personal blocks first, and events can only book into the right one.
  • Made setup a wizard, not a settings page, with tap-to-place, color-coded blocks finished in one sitting.
  • Let the template say no, so declining a “free” slot felt like policy, not guilt.

Results

Under 5 minutes 4 of 4 testers Faster decisions
Wizard setup Completed it Across three conflict scenarios

One tester summed it up: “It’s not me saying no, it’s my schedule saying that time is already allocated.”

Custom Web App

Contour Skill Visualizer

Designed and built a self-assessment tool that turns a designer's ranked skills into a visual shape (I, T, Pi, M, or X), with a growth guide and exportable charts. Solo concept, UX, motion, and front end, shipped in three weeks.

Every designer says they’re T-shaped. I built a tool that makes you prove it: pick your skills, rank them, and see whether you’re actually I-, T-, Pi-, M-, or X-shaped.

What I did

  • Made honest answers the only option. Only one skill can be a 10, one a 9, and one an 8, so you can’t call everything “expert.” That one rule is what makes the shape mean something.
  • Scored the shape instead of eyeballing it. The tool reads your ratings as a profile: how many skills sit at 8 or above, how far your top skill stands over the next, how many distinct peaks rise out of the chart, and how evenly high the whole thing is. Each of the five shapes earns points from those signals. One standout peak leans I, two well-separated peaks lean Pi, three or more lean M, and a high, even plateau leans X. The highest score wins, and T is the fallback when nothing stands out.
  • Tested accessibility instead of assuming it. Playwright and axe check keyboard and screen-reader behavior, tooltips work on focus and tap, and every animation respects reduced motion.
  • Sweated the motion details. The download icon redraws itself stroke by stroke on hover. Three rounds of prompting only made it blink. Re-exporting the icon as strokes instead of a filled shape is what made the draw-in possible.

Results

5 shapes 3 export formats 0 answers stored
I, T, Pi, M, and X PNG, JPEG, SVG Private by design

The lesson from the icon: when you build with AI, the source file is the spec.

Custom Web App

LinkedWin Content Tool

Designed and built a personal LinkedIn content dashboard covering writing, scheduling, a searchable post library, and analytics. It went from an empty Next.js project to production in about 53 hours and merges four disconnected data sources into one library.

I needed one place to write, schedule, and look back on my LinkedIn posts. Free tools only did one of those jobs, so I designed and built LinkedWin over a weekend.

What I built

  • A composer built around how I write. Its insert bar holds the symbols and structures I use most, plus saved footers, @mentions, and multiple pinned comments.
  • A post library that reconciles four data sources. LinkedIn and Favikon use different IDs for the same post, so the tool matches them by caption and date. Estimated numbers are labeled as estimates.
  • One shared database for every host. Local SQLite split production data between Vercel and the worker, so I moved both onto Turso and made silent failures throw errors instead.

Results

~53 hours 7 sections 702 posts
Empty project to production Overview, Calendar, Composer, Library, Data, Radar, and Engage Four sources, one library

The auto-repost bot works, but it stays off. A banned LinkedIn account would cost more than any bot saves.