At Roboception, I develop, sell, and lead. My company deploys me on trade show floors all day because I sell. I generate the highest lead counts (38 leads at Automatica on day one) because I understand both deep technical detail and what customers actually need to hear. I maintain the most-used public codebases. I lead cross-company technical collaborations. I forge real partnerships.
I designed and shipped GRI—a standardized integration layer between industrial robots and Roboception's vision systems—into production firmware globally (shipped in 25.10 release). It works across KUKA, ABB, Fanuc, Universal Robots, Techman. Started as a free Docker container given to partner companies for feedback, iterated based on real users, shipped production version.
I maintain Roboception's public KUKA robot codebase. When engineers Google for KUKA integration examples, they find my repository. It's become the reference integrators rely on globally.
At RoX hackathons (Intrinsic at Google Munich, VDMA Frankfurt), I shipped the first end-to-end digital asset flow through T-Systems' data intelligence hub; leveraged partner credibility to force critical infrastructure specifications. Built trust with people across companies—three remain close collaborators.
First DIH asset exchangeon GitHubRobot programmer at 11 PM loosing his mind on Christmas Video (sound)
Collaboration between Roboception and Intrinsic for Automatica 2025 RoX AI
Roboception Logimat 2025 Demo. CAD, assembly, and programming
I got RSI from constant typing. Needed a transcription solution better than anything on the market. Built VibeVoice from scratch.
Lightning-fast voice transcription at 3× the speed of typing. Built while working full-time at Roboception and moving apartments. Competing on speed and price in a crowded market.
                    662 verified users in just 2 months. Started as Vibecoding for developers with RSI, now the real TAM: anyone who types.
Building native cross-platform in Rust is hard: audio APIs, global hotkeys, clipboard paste—all platform-specific. Real-time streaming, sub-second partials across Windows, macOS, Linux. Solution: Ship early, give away free, iterate with users. I gave out free memberships to everyone I know, collected feedback, moved fast. Real users teach you what matters.
This isn't just transcription. It's the foundation for ambient AI assistance. If my aunt can speak her needs and have systems respond—medication, navigation, emergency detection—voice input must be instant, reliable, universal.
From PCB design to full-stack software to closing deals on trade show floors
From CAD to code to deployment. Hardware, software, and integrations that solve real problems.
                        
                    Automated medication with HA/MQTT. Schedules supplement timing windows. Building block for my aunt's medication management.
Amazon removed third-party Shopping List integrations. I built the reliable bridge into Home Assistant — region-aware Email+OTP, battle-tested. I talk to my house. It listens.
Learning from Demonstration framework for instant, code-free robot teaching, verified in Siemens production. Parallel development: specialized gripper for flexible metal sheet transfer.
                        
                    Low-power plant monitoring with auto-watering. Custom PCBs. Ambient sensing and automated response at scale.
I've proven I can build products people pay for (662 users in 2 months).
I've proven I can ship production systems (GRI shipped in 25.10 release).
I've proven I can sell (38 leads, day one).
I've proven I can lead (RoX standardization across companies).
But every month on German bureaucracy is a month NOT shipping features users need.
I'm impatient for impact. I want to build where speed and ambition are advantages.
Ready to build the AI-assisted future together.