UX/UI for AI-native, workflow-heavy products.
Selected case studies · 2023–2026
How I make UX/UI decisions
I focus on how work moves through the system, who is involved, where risks appear, and what should be automated or kept human. I model information architecture, states, permissions, decision points, and orchestration rules so the product is clear, auditable, and ready to ship. The design system is where that model becomes reusable across a product instead of solved once per screen.
I consider delivery constraints early, including team capacity and rollout plan, ensuring the design is realistic and ready for future growth. My goal is to improve activation, adoption, and retention — not just usability — and shape systems that remain effective as the product evolves. On Strata, this discipline extended to a B2B2C layer — I designed the checkout and tracking flow for client-facing sales pages, delivered as wireframes and a front-end prototype, with a deterministic customer-facing boundary.
A design system has three layers, and only one of them says anything about how a designer thinks: the primitives are table stakes, the components are a catalogue, and the layer above is where judgment lives — uncertainty, provenance, generation state. That last layer is what I published at /storybook, because it is the one no library ships and the one that decides whether people trust an AI feature.
Build systems that work in practice, align teams, and support real product execution. I design AI workflows and UX for enterprise products that reduce manual work, lower cognitive load, and keep humans in control. I consider current operations and future needs so teams ship relevant solutions. The design system turns those workflow states and decisions into reusable rules — semantic tokens and components with defined states and behavior — shipped as a published Storybook the frontend team pulls directly into its own workflow.
Strata and Repfabric are the two proofs of that discipline, and they pull in different directions on purpose. Strata turned fragmented manual operations into a System of Action — states, roles, and an AI layer that acts only where the decision surface is narrow enough to trust, with anything past that designed and deliberately held back until audit trails and permission scopes are safe. Repfabric turned a multi-role CRM/ERP into a single, deliberately scoped experience built around the one persona who actually drives adoption. Same capability in both: turning operational complexity into decisions and systems a team can ship.
Three of the case studies below come from two client engagements. Strata and the design system behind it are the same one: Senior UX/UI Engineer, 2024 — Present, with Avanto, via Agentic Dream, on a platform whose enterprise accounts include CBRE and Bank of America. Repfabric is the other: Senior UX/UI Engineer, 2023 — 2025, on a CRM/ERP platform with a designed Salesforce data-sync mapping — the mapping and how it would be incorporated were mine to design; an engineer implemented the integration into the live build. The fourth is my own product: the assistant on this site, which I designed, built and deployed on my own.