I turn design into
organizational advantage.

Kroger's personalization and discovery systems touch nearly everything a customer finds and buys online. When I joined, design had almost no influence over how those systems got built. Data science owned the roadmap. Engineering owned the execution. Design got called in at the end to make things look reasonable — which meant the hardest problems had already been decided without us.

I stopped waiting to be invited and started showing up earlier. That meant getting into the rooms where ML models were being scoped — not to approve them, but to ask the user questions nobody else was asking. What happens when the recommendation is wrong? What does the customer see, and what do they do next? Rapid prototyping became our way of pressure-testing algorithmic decisions before engineering spent weeks building something that confused people. We used LUMA facilitation methods to run discovery sessions that got design, product, and data science speaking the same language.

We shipped Kroger's first conversational AI shopping tool. The ML Ops platform work cut recommendation deployment cycles from weeks to hours. The 'Yearly Checkout' feature, a personalized annual summary, drove real gains in retention. The harder work was helping engineers and data scientists understand what design actually contributes before the decisions get made. That one doesn't show up in a release note.

Delta's internal tools, crew scheduling, turbulence tracking, command center operations, had been built by different teams at different times. Nobody designed them to work together because nobody was ever asked to. When we talked to ops staff, one phrase kept coming up: situational awareness. The ability to know exactly what's happening, right now, across all the moving parts. Context-switching between disconnected systems with their own logic and terminology was chipping away at that. Under normal conditions, that's friction. Under irregular operations, it's a serious problem.

The client came in asking for a UI refresh. Before committing to a direction, we spent time with operations staff watching how they actually worked: what they checked first, where they got stuck, what they'd learned to work around. That shifted the conversation. A fresh coat of paint on four disconnected systems wasn't going to restore situational awareness. Working together with the client, we reframed it: what the team actually needed was a shared design language that let someone move between tools without losing their footing.

The design language we built got adopted org-wide — not as a style guide people reference once and forget, but as the actual standard new tools were built to. Getting teams who own their own products to share a common foundation takes more convincing than any deliverable. We got there. Took a while.

Connected TV and OTT streaming were still relatively new territory for sports media. The NBA had web and mobile figured out, but integrating streaming platforms into a coherent fan experience was a different problem. Each surface had its own content model, its own interaction patterns, its own assumptions about what a fan needed in that moment. The question wasn't just how to make things look consistent. It was how to build something intentional across all of it, starting with the newest and least understood piece: OTT.

We built a framework, not a design system, but a decision-making tool. It mapped how fan context changes across surfaces and how those surfaces talk to each other. In a multi-screen experience, the TV isn't just a big phone. The phone isn't just a small TV. Someone watching a game on their couch with a phone in their hand is doing two things at once, and the experience should account for that relationship, not just each screen in isolation. The framework gave teams a way to make those decisions intentionally. We ran working sessions with stakeholders across digital, broadcast, and arena operations to test it against real scenarios before it went anywhere near a product.

The framework got adopted across NBA digital. But the more useful thing it changed was what the internal conversations were even about. Less "why doesn't mobile match web" and more "what does someone actually need in this moment, on this screen." That's a better conversation to be having, and it was a new one for them.

84.51° is Kroger's data science and insights arm, smart people building analytics products for large enterprise clients. Each product team had their own designers, their own component libraries, their own way of solving the same problems. That works fine until a client uses two products and wonders why they feel like they came from different companies. It also slows everyone down, because every team is solving problems the org has already solved elsewhere.

Before pushing for a design system, which is the obvious answer and sometimes the wrong one, we mapped where the actual pain was. Where were users confused? Where were designers rebuilding the same thing twice? That gave us somewhere specific to start. We built around those gaps and kept the teams involved so the whole thing didn't show up one day as a mandate nobody asked for.

Teams adopted it and kept shipping, which is honestly the only metric that matters for a design system. If it slows people down, it gets ignored. This one didn't. Client feedback on consistency improved, and designers stopped spending half their time rebuilding things that already existed somewhere else in the org.