Are you happy now? You just destroyed my whole page.
I worked really hard on this, you know…
*sigh* Can you at least pass me the broom? 🧹
Full-stack with a backend & infrastructure heart. Ten-plus years shipping web platforms, microservices and Kubernetes clusters — from Berlin, now from Bangkok.
I've been building websites and microservices for more than a decade. Most of my experience is full-stack, but my centre of gravity has always been the backend — and lately, the platform underneath it.
I started with a German dual vocational programme (two weeks in the company, one week in school), was a fully qualified team member soon after, and three years later took a team-lead role to learn everything outside the editor: 1on1s, OKRs, stakeholder alignment, and translating requirements between systems, teams and external contractors. Scrum and Kanban were part of every role.
What I care about: developer experience (automated formatting, static analysis, CI that tells the truth), boring reliable infrastructure, and clear ownership. I'm not easily rattled by job stress, I ask questions early, and I enjoy sharing what I know — with people from different roles, technical and cultural backgrounds.
Off the keyboard I've got a media-design background (video and audio), which still shows whenever there's a Photoshop layer or a stream overlay to fix.
Dots are a gut feeling, not a certificate. Five = I've run it in production for years and would pick it again.
Things I'm working on right now — or shipped recently.
A supply-chain visibility platform for freight forwarders — live shipment tracking across sea, air and road, a live map with real ocean routes, a risk radar that matches world events against live cargo, and CO₂ reporting. I build the backend: Node/TypeScript services on MongoDB, ScyllaDB and MySQL, running on Kubernetes.
A cold-chain project at Neolink: temperature-controlled boxes for sensitive goods like pharmaceuticals, travelling the world with cellular IoT trackers inside. I designed the delivery detection — a three-fence geofence state machine (origin → approach → delivered), computed server-side from device webhooks, ending in an automatic cold-chain certificate.
Moving Neolink from one shared stack to per-customer infrastructure: a central admin cluster runs ArgoCD, Vault, monitoring and identity, and every customer gets their own Kubernetes cluster and databases — described in a single file in Git. Onboarding a customer or shipping a new app is just a pull request.
Carrier aggregators confirmed deliveries over an hour late for long-tail regional couriers — and meant asking customers for tracking numbers.
The IoT tracker inside the box always knows where it is. We made the device the sole delivery signal, evaluated server-side with geofences.
Geo lookups, e-mails and certificates inside the webhook handler risk timeouts — and the sender then marks the delivery as failed and retries.
Validate, persist, enqueue, return 200. A worker does the heavy lifting — and brings retries, dead-letter and replay for free.
All customers on one shared stack: unclear per-customer cost, shared blast radius, and AI features risk cross-customer data exposure.
One cluster and one set of databases per customer, defined in a single Git file — isolation, cost visibility and a sales argument in one.
Coding agents are fast — but happily skip planning and touch things they shouldn't.
Pre-tool hooks and work-guards gate every step: plan first, then code. Fast, but on rails.
Backend and infrastructure lead at Neolink, a supply-chain visibility platform for freight forwarders. I build the shipment-tracking backend, designed the geofence-based delivery detection for the BeeBox360 cold-chain project, and am moving our stack to per-customer Kubernetes clusters on DigitalOcean, deployed via GitOps.
Symfony backends, WordPress and Google Maps integrations for agency clients, plus first responder for customer emergencies. Introduced automated code formatting, developer 1on1s and testing tools, always looking for ways to improve developer experience.
Led the web team while staying hands-on: regular 1on1s, team-lead syncs, and close work with stakeholders across departments to hit company OKRs.
New and existing microservices, A/B tests, bug fixing, dependency upgrades — Jira, weekly planning, daily stand-ups. Also kept the legacy Zend Framework 1 shops alive.
German dual system: two weeks in the company, one week at school, final IHK exam passed.