Forward Deployed AI / UX Engineer · Lumoview
2021 – today · Cologne
Lumoview builds scan-to-BIM products for capturing, managing and digitally reconstructing buildings. I was the only UX/UI designer working across all three product surfaces — LumoApp, LumoCloud and LumoCad — and also implemented product features in React, React Native and TypeScript.
- Worked directly with surveyors, modelers, quality checkers and domain experts; turned observations and conversations into clear workflows rather than passing requirements through layers of abstraction.
- Led UX research, information architecture, interaction design, interface design and design-system work across mobile capture, cloud management and browser CAD.
- Used empathy and domain understanding to reduce dense technical workflows to their essential decisions, while keeping expert control where it mattered.
- Replaced static hand-offs with AI-assisted code prototypes, moving from a user problem to something testable in hours instead of waiting for a full implementation cycle.
- Used Claude Code, Cursor and context-engineered agent workflows throughout implementation. More than a hundred commits in the available product histories carry Claude or Cursor co-authoring trailers.
- Built production frontend and mobile features, including offline-first data flows, device/scanner interactions, 3D/CAD interfaces, comments, measurements, progress, quality control and multilingual field workflows.
- Contributed to APIs, automated tests and delivery pipelines including GitLab CI, GitHub Actions and Xcode Cloud.
- Interactive prototyping — Research, design and implementation collapsed into one AI-assisted loop: working product prototypes instead of static hand-offs.
- LumoApp — Lumoview's offline-first mobile field app for turning buildings into complete, structured scan projects.
- LumoCad — A browser-based 3D CAD editor that turns scan data into structured building models and deliverables.
- LumoCloud — The web platform around a building scan — projects, buildings, data, files, teams, quality control and hand-off into LumoCad.
- Multi-scan merge — An agentically developed Python pipeline for registering multiple scans, scoring merge quality and grouping rooms.
- pano-app — An early browser prototype for exploring building scans as connected 360° panoramas.