Client: Docker
Timeline: Neverending
Roles: Lead Product Designer, Visual Design
Prototypes:
Docker Hub is old, and it shows. The interface looks dated, the information architecture is messy, and years of different teams bolting on features have left it feeling disjointed. There have been plenty of attempts to fix it, but none of them ever really stuck.
My push to change that started during a company-wide hackathon with a few teammates. I took what we already had on Hub and pushed it a bit further, nothing drastic, just a few thoughtful steps into the future. We ended up winning the hackathon, and the work received positive feedback from leadership.
That momentum led to the formation of a Visual Design Tiger Team. The goal was to completely rethink Docker’s visual language as if nothing existed yet and start with Hub as the first surface before expanding to the rest of Docker’s products. Those sessions were full of energy, brainstorming, defining design principles, building stylescapes, and merging them into a unified direction. Things were moving fast and looking promising until, like many good initiatives, it fizzled out before it could fully take hold.
It picked back up after a leadership change. The new executive team wanted Design to level up Docker’s design quality and direction.
The new executive team came in and wanted Docker to raise the bar for design across the company. With my manager on maternity leave, the acting manager started a new team to focus on a North Star Vision. Early on, it felt more like cleanup work than a true vision. The focus was on fixing inconsistencies instead of defining where Hub should go. I joined in anyway to help where I could.
Things got more complicated after the launch of the MCP Marketplace. Leadership loved how it looked and wanted Hub to match it. I understood why; it looked good, but it turned our vision work into a simple visual refresh. No real problem-solving, no direction, just new paint. I raised concerns with my temporary manager, who dismissed them, so I reached out to other designers and managers to sanity-check my thinking. They all saw the same issues.
When one of the managers encouraged me to keep pushing the vision work and focus on the kind of work my actual manager, the Head of Design, would want to see when she returned, I decided to take it further.
I stripped everything back to the essentials and started fresh. The new mockups leaned into a modern, minimal aesthetic inspired by trends across the developer tooling space. Products like GitHub, Hugging Face, and Vercel have all moved toward clean, content-first interfaces with generous spacing, muted color palettes, and typography that creates hierarchy without adding visual noise. I wanted Docker Hub to feel like it belonged in that same space—simple, confident, and calm.
From there, I explored how AI could make the experience feel more adaptive and intelligent. I started asking, what if Hub could surface the right content at the right time? What if search and discovery were guided by context, not just keywords? These explorations weren’t just visual—they tied directly back to real user problems: the overwhelming amount of content, poor discoverability, and the lack of distinction between premium and free content.
Once the direction started to feel cohesive, I shared it across the design org and with a few key leaders. The response was energizing and helped reignite excitement around what Docker Hub could become, setting the stage for how we think about its future today.
Not long after, I held a group brainstorm where the design team was brutally honest about what wasn’t working in Hub.
That same week, I ran a mini design sprint with ICs and leaders working directly on Hub.
My manager is now back from maternity leave, and the work has picked back up again. The vision work we started last year has become a major driver behind the new Docker Design Roadmap, where every team will be exploring blue-sky concepts in parallel. The Hub work is leading that effort, serving as the tip of the spear and helping set the direction for how we think about design across the broader Docker ecosystem.
Vision work seems to be hard at Docker, or maybe everywhere?
While this work was happening, Docker was also making its big push into the security space with Hardened Images, which are hosted on Hub. That connection gave me a way to tie real business impact to the vision work. The reality is that Hub doesn’t look or feel secure, and that matters when our newest products are built around trust. I’ve started using this narrative when speaking with leadership, and it’s beginning to resonate.
As for the dissmisive pushback I got from my acting manager, I’ve had time to reflect. I don’t think I would have done anything differently. It was clear what needed to be done and just as clear that we were heading in the wrong direction. I’ll admit, the pushback caused a fair amount of anxiety and self-doubt, but the support I received from other managers and the acting head of design helped me stay grounded and keep moving forward.