Work

Designing and building the user-facing layer of an enterprise knowledge-graph platform — where the core challenge is making sophisticated AI and semantic data capabilities legible to domain experts who aren't data engineers. Work spans interaction design, UI architecture, and LLM integration, with a consistent focus on keeping users in control of complex, high-stakes knowledge.

Four years focused on a single, hard question: how do you design AI that domain experts actually adopt and trust? I built knowledge graph tools for procurement specialists at Bosch, ran structured user studies, measured changes in task performance and decision quality, and iterated until the AI fit the workflow — not the other way around. The answer required equal parts design research, AI engineering, and organizational understanding.

Built a digital procurement platform from the ground up for hundreds of internal users. Starting from blank architecture, I drove the frontend through design, implementation, and handoff to a team of engineers — keeping the interface coherent and usable as complexity scaled. The platform handled sourcing and supplier workflows that previously lived in email threads and spreadsheets.