Resume & Career Timeline

The layoff ended a run of over 11 years at Hearst. Rather than sprint back into the same market, I made a bet: the designers who matter in the next decade will understand AI systems the way I’d learned to understand front-end code, from the inside. Most people hype LLMs from the outside; my position is different: a design creative with real footing in coding, devops, and infrastructure/virtualization. So I built the stack from the metal up: virtualized GPU inference in my homelab, a cost-routing proxy I wrote myself, agent pipelines I supervise like a design team. No black boxes, nothing I can’t explain. That practice is now the medium I design in.

The bet was asymmetric, which is the only kind worth making. The downside was bounded: a few months, a $20/month plan, and a 16GB GPU that was just sitting there — COVID was over, I didn’t have time to play Call of Duty, and 2023 became unplayable. The upside was the entire next decade of my career.

*How this was measured. The 95%+ figure is a Deque Axe Monitor score from automated scanning. It is not a manual full-conformance audit and it is not a certification. WCAG is a W3C standard and conformance is self-declared, so no US body certifies it. Scans covered first-party content the team controlled and excluded third-party embeds such as ad networks, where responsibility for defects is contested. The 85 core sites are what was scored. The 120+ figure is the wider footprint that was audited.

Why this one stuck with me

Design systems work is uncredited by design. You build the thing that lets other people’s names go on the page, and the better it works the less visible you are. Over eleven years the platform carried the bylines of thousands of writers, editors, and photographers, and none of that surface said who built the surface.

The masthead is the exception. For one month in early 2014, before the platform work took over, I was embedded with the Esquire editorial team learning how they actually worked. The credit is a record of that period rather than of the decade that followed.

What I Learned

Editorial teams are a unique breed of users. They need powerful tools that can handle complex content, but they also need those tools to be intuitive and easy to use. If you don’t build the right tools, they’ll hack together their own solutions using what’s available. This can lead to a lot of technical debt and security issues. I learned that building a CMS is not just about building a tool for editors, it’s about building a tool that can scale and evolve with the needs of the business.

What I Learned

Wearing both the product and design hat in a small team forced me to think holistically about every decision. Design choices had to be justified with data, and product decisions had to be grounded in user experience. A/B testing became my tool for resolving subjective debates with objective evidence: a habit I’ve carried forward ever since.