I lead software engineering at devolo, where my team builds software for millions of connected networking devices. My background spans IoT infrastructure, mobile, web, HCI research, and regulatory compliance — I pick up new domains quickly, which helps when the problems keep changing.
I care equally about the people side: setting goals that engineers actually understand, supporting growth, and creating space for strong people to do their best work. I’ve supported multiple engineers through promotions and role transitions, and worked with VP Engineering and VP Product to define team-level goals that give everyone clarity on what success looks like.
If this resonates, let’s connect — I’d love to hear from you.
Previously: Expert SWE (05/2021–08/2023) · Product Manager (09/2023–08/2024)
Proposed and implemented contextual in-app rating prompts at high-satisfaction moments (installation complete, dashboard, speed test). Combined with UI widget contributions, raised app store ratings from ~3.5 to 4.2 across 50,000 active users over six months.
Designed and built a cross-platform replacement for a Windows-only Access-database compliance process with no concurrent editing. Covers the full current product family and serves as the foundation for all future product lines.
Built a MkDocs + EuroPDF pipeline replacing a single-author Adobe workflow. Turnaround went from months to days; developers now contribute directly; PDF/UA-1 accessibility compliance is automated; three products shipped so far.
Sole developer and maintainer of devolo's fork of Eclipse hawkBit, managing OTA firmware rollouts across 4 million connected networking devices. Diagnosed a recurring database bottleneck — a monitoring service running full-table SQL scans every 30 minutes was holding DB CPU in the high 90s and causing UI freezes. After the fix, CPU settled around 60% and freezes stopped. Built automated cleanup schedulers for rollout records and action status data, delivering a sustained ~7% performance improvement. Also contributed fleet observability via a custom telemetry endpoint, dynamic poll-interval overrides, and managed three upstream version migrations.
Two peer-reviewed, open-source tools built during PhD research at RWTH Aachen University. TRACTUS (CHI 2020 full paper) is an RStudio extension that detects and tracks code experiments in hypothesis-driven data science — helping researchers retrace analytical decisions and avoid redundant work. StatWire (CHI 2018) is a hybrid IDE for R that adds visual data-flow editing alongside text, making statistical pipelines easier to understand and replicate. Both tools have been cited and adopted by researchers outside the lab.
Building a replacement for an underperforming support chatbot that prioritises accurate "I don't know" responses over confident wrong answers. In testing — baseline accuracy benchmarks will be captured before deployment.
Led devolo's first internal AI coding assistant session (March 2026), attended by 5 people outside engineering. Direct outcome: 12 GitHub Copilot seats with raised spending limits. Running sessions on a two-month cadence.
Download the PDF for the full CV.