Backend & systems engineer — Germany

I build the backends that carry millions of users.

I'm Jerome — a systems engineer and CTO. For five years I've designed distributed backends, shipped mobile and marketplace products, and built the platforms other developers rely on. I started young and I've been in production ever since.

See the work

Currently — CTO at Valorbyte & Shabo Group

fig.00 — distributed systems
Role
Systems engineer / CTO
Based
Rügen, Germany
Coord
54.42°N · 13.43°E
Building since
2020
Users served
4.4M+

Four systems I've lived inside

Four systems I've spent real time inside — what they were, what made them hard, and what I'd tell myself before starting again.

Flagship
01 / 04

GommeHDnet

Java Developer

Period
2020 – 2024
Scale
4.4M+ registered players
  • Java
  • Netty
  • Cassandra
  • Redis
  • MySQL
  • ZooKeeper

Game frameworks for a network of 4.4M players

Shared frameworks and mini-games running across one of the largest German multiplayer networks — code where a single dropped tick is visible to thousands of people at once.

Context

GommeHD is one of the biggest German-speaking multiplayer networks. I worked on the mini-games and the internal frameworks underneath them: matchmaking, game state, player data, and the plumbing that connects dozens of game servers.

The hard part

Everything is concurrent and nothing forgives latency. A player joining a round, a stat being written, a server rejoining the cluster — all of it happens thousands of times a minute, and any stall is something real people feel instantly. And the frameworks weren't only mine: other developers built their games on them, so an awkward API was a cost multiplied across the whole team.

Approach
  • Treated the framework as a product with internal users — clear boundaries, sane defaults, and APIs that were hard to misuse.
  • Kept hot paths off the database: Redis for volatile state, Cassandra for write-heavy player data, MySQL where relational integrity actually mattered.
  • Leaned on ZooKeeper for coordination and Netty for transport, so servers could come and go without taking live rounds down with them.
Outcome

Games and frameworks that held up under the network's real traffic for years — used daily by millions of players, and by the developers building on top of them.

What it taught me

Writing infrastructure other engineers depend on taught me more than any single feature could. The API you ship is the one people live with, so I learned to optimize for the second reader of the code, not the first writer.

02 / 04

Shabo Group

Chief Technology Officer

Period
2025 – present
Scale
Multi-tenant, franchise-wide
  • Java
  • Spring
  • TypeScript
  • React
  • PostgreSQL

A marketplace for local Edeka stores

An online-ordering platform that lets independently-run franchise stores sell locally without each one reinventing a webshop.

Context

Local Edeka stores run as independent franchises. Each needs an online storefront, but none should have to build or operate one alone. We are building the shared platform: one system, many stores, each with its own catalogue, pricing and logistics.

The hard part

Multi-tenancy is the whole problem in disguise. Stores need to feel independent — their own products, prices, delivery zones — while everything runs on one codebase and one operations team. Get the tenant boundaries wrong early and you pay for it in every feature afterwards.

Approach
  • Modeled the store as the tenant boundary from day one, so isolation is a property of the data model rather than something bolted on later.
  • Kept ordering, catalogue and logistics as separate domains, so each can evolve — and be reasoned about — on its own.
  • Chose boring, well-understood tools so the team spends its attention on the domain, not on the framework.
Outcome

A platform taking shape that store owners can run without a technical background, and that we can extend to new stores without forking the code.

What it taught me

As CTO, the hardest calls are the ones you make before there's traffic to justify them. I spend architectural budget where reversing the decision later would be expensive, and stay cheap everywhere else.

03 / 04

Valorbyte

CTO & Managing Director

Period
2025 – present
Scale
Mobile games + platform
  • Java
  • Spring
  • Unity
  • C#
  • ASP.NET

A platform for skill-based competition

Mobile games, plus the backend that turns them into fair, real-money skill competitions.

Context

Valorbyte builds mobile games and a platform that runs skill-based competitions on top of them. The games are the fun part; the platform is where trust lives — matchmaking, results, payouts, and the rules that keep it fair.

The hard part

When money rides on a match, correctness and fairness stop being nice-to-haves. Results have to be verifiable, matches have to be tamper-resistant, and the whole thing has to feel instant on a phone with a flaky connection.

Approach
  • Split the playful client from the authoritative backend: the game entertains, the server decides.
  • Designed the competition and wallet logic to be auditable, so a disputed result has a single source of truth.
  • Built the platform to be game-agnostic, so a new title plugs into matchmaking and payouts instead of rebuilding them.
Outcome

A shared platform new games can sit on top of, rather than a stack of one-off backends.

What it taught me

Running engineering here means I own the trade-offs end to end — from a Unity frame budget to the company's cash flow. It's made me far more deliberate about what 'done' actually means.

04 / 04

Independent

Personal project

Period
Ongoing
Scale
Research & prototype
  • Rust
  • WASM
  • Kubernetes
  • Hetzner

Serverless functions on German infrastructure

A function platform built around data sovereignty: deploy small units of code that run entirely on German-hosted infrastructure.

Context

Plenty of European teams want serverless ergonomics without their data leaving the jurisdiction. This is my ongoing exploration of what that platform looks like — push a function, get an endpoint, and know exactly where it runs.

The hard part

Serverless is easy to describe and hard to make cheap. Cold starts, isolation between tenants, and predictable performance all fight each other — especially without a hyperscaler’s scale to hide the costs.

Approach
  • Chose WASM as the isolation unit — small, fast to start, sandboxed by design — with Rust for the runtime around it.
  • Targeted commodity German hosting to prove the economics work without hyperscaler pricing.
  • Kept the developer surface tiny: a platform’s job is to disappear.
Outcome

A working answer to a question I kept running into: can you have serverless ergonomics and data sovereignty at the same time?

What it taught me

Building this from scratch is the clearest reminder of how much a platform does for you invisibly. Every convenience I’d taken for granted is a decision someone once made — and now I’m the one making them.

Who's behind the systems

I got into this the practical way. My first paid work was keeping a company’s website running and, on one memorable occasion, programming a Siemens controller for an aquaponics system in ladder logic — a sentence I still enjoy saying out loud.

From there I spent four years building game frameworks for a multiplayer network used by millions of people. That’s where I learned what “at scale” actually costs, and why the unglamorous parts — data models, coordination, APIs other people rely on — are the parts that decide whether a system survives.

Now I lead engineering as CTO at a few companies: a skill-gaming platform, a franchise marketplace. So I spend about as much time on architecture and helping a team ship as I do writing code myself. What I care about is the boring kind of good — systems that stay up, code the next person can read, and decisions that still look sensible a year later.

Tools I actually reach for

No logo wall. This is the working set — grouped by where in a system it lives.

Languages
  • Java
  • Kotlin
  • TypeScript
  • JavaScript
  • C#
  • Rust
  • Python
Backend
  • Spring Boot
  • Spring Cloud
  • ASP.NET
  • Axum
  • Actix-web
  • REST
  • gRPC
  • WebSockets
DevOps & Cloud
  • Docker
  • Kubernetes
  • CI/CD
  • ArgoCD
  • Terraform
  • AWS
  • Google Cloud
  • Hetzner
Data & Messaging
  • PostgreSQL
  • MySQL
  • MariaDB
  • MongoDB
  • Cassandra
  • Redis
  • Kafka
  • RabbitMQ
Frontend & Mobile
  • React
  • React Native
  • Angular
  • Astro
  • Flutter
  • HTML/CSS

How I got here

Not a straight line. A web presence and a PLC, four years of games at scale, then leading engineering — with a few honest day jobs along the way.

  1. 12/2025 - Present

    Partner

    — Wolff-Consulting GbRnow
  2. 07/2025 - Present

    Managing Director

    — WIG Capital UG (haftungsbeschränkt)now
  3. 06/2025 - Present

    Chief Technology Officer

    — Shabo Group GmbHnow

    Leading of the development of a marketplace for online orders at local Edeka stores via a franchise system.

  4. 03/2025 - Present

    Chief Technology Officer & Managing Director

    — Valorbyte UG (haftungsbeschränkt)now

    Leading the development of mobile games and a gaming platform for skill-based competitions

  5. 10/2020 - 09/2021 & 09/2022 - 03/2024

    Java Developer

    — GommeHDnet GmbH

    Development of mini-games and frameworks for a multiplayer network with over 4.4 million registered users, using modern technologies such as Cassandra/Datastax Enterprise, MySQL, Redis, Zookeeper, the client-server framework Netty, Maven as the build tool, and Git as the version control system.

  6. 06/2023 - 11/2023

    Customer Service Representative

    — regiocom Customer Care SE

    Handling customer inquiries via phone, email, contact form, or (WhatsApp or web) chat, providing information on billing questions, products, prices, tariffs, or contract data, order processing for an electricity or gas supply contract, recording cancellations, revocations, powers of attorney, and death notifications, and documenting concerns in the database.

  7. 09/2021 - 10/2022

    Executive Assistant

    — Rügentest

    Conducting and monitoring SARS-CoV-2 PoC antigen tests, swab sampling and documentation of SARS-CoV-2 PCR tests, maintaining customer data, setting up hardware and software, customer support via phone and email, personnel management, and assisting the management.

  8. 06/2021 - 09/2021

    IT Specialist

    — KFR Kreide-Farbenwerk Rügen GmbH

    Maintenance and upkeep of the web presence for affiliated companies, mobile app development with a focus on React Native and TypeScript in the frontend, as well as Java, Spring, and MariaDB in the backend, programming a Siemens PLC (Programmable Logic Controller) for use in an aquaponic system using ladder logic, function block diagram, instruction list, and structured text.