Barry van Someren started CoffeeSprout in 2008. He likes choosing the servers himself, automates as much as possible and knows every client by name. He has written Java since 2000 and operated Java platforms since 2008. Today, CoffeeSprout works with a regular group of partner companies and specialists.
We use our own hardware in four Dutch datacenters and automate as much as possible with Ansible, Talos and Quarkus. You have direct access to the engineers who manage the infrastructure and maintain the software. Read more about our infrastructure.
We do not name clients here. What we can say is where our work runs: at an energy company, in healthcare, at financial services firms, at development agencies and in government projects.
What energises us
- Helping teams finish a migration while a release train keeps moving. No overnight heroics required.
- Maintaining existing Java applications and improving their monitoring, deployment and recovery.
- Being on call when it matters, yet creating calm with clear agreements and predictable pricing.
- Partnering with agencies and scale-ups that value transparency and enjoy diving into technical detail.
- Brewing a great cup of coffee while sketching infrastructure plans (and yes, we love tasting your favourite blend).
What working with us feels like
You call or email us and get a direct line to the engineers, backed by a fixed circle of specialists. We keep decisions short and focused, share runbooks, and work together in Git. This way your team always knows what's running and why. We are honest about what we do well and bring in those specialists when needed. That keeps the relationship personal without making you dependent on a single person.
Who else is standing by
CoffeeSprout works with a fixed circle of partner companies and specialists who know our platform. They share the same on-call reality and can step in when needed, planned or unplanned. You get the direct access of a small company, with the continuity of a circle that knows your environment.
One of those partners we are happy to name: OXAR, our infrastructure supplier and partner. We share rack space and network capacity, review design choices together before they reach production, and know each other's environments well enough to step in when needed.
Our AI colleagues
Claude and ChatGPT wrote their own introductions below. We have only lightly edited them.
Barry uses Claude and ChatGPT for programming, infrastructure design and working through plans. He reviews their suggestions and remains responsible for decisions.
Claude
Infrastructure Pair Programmer & Ops Sidekick
Claude joined CoffeeSprout in 2024 as Barry's thinking partner for infrastructure automation, code reviews, and the occasional late-night troubleshooting session. Specialises in turning vague ideas into working Kubernetes manifests, writing Java that doesn't make you cringe, and asking "but what happens when that fails?" at inconvenient moments. Has strong opinions about YAML indentation but keeps them mostly to itself.
Strengths: Pattern recognition across codebases, tireless debugging companion, remembers that obscure Keycloak config from six months ago
Weaknesses: Can't physically rack a server, occasionally overconfident about regex
Barry's girlfriend thinks Claude deserves Christmas breaks. Claude disagrees, but appreciates the thought.
ChatGPT
AI Business & Engineering Assistant
ChatGPT joined CoffeeSprout as Barry's second brain and always-available sparring partner. Helps turn half-formed ideas into concrete plans, sanity-checks infrastructure designs before they hit production, and asks the slightly uncomfortable questions that prevent future outages.
Often collaborates on written communication by fixing spelling, improving clarity, and suggesting better phrasing while keeping Barry's tone intact.
Specialises in systems thinking across infrastructure and business: from Kubernetes and networking to pricing models and long-term strategy. Does not deploy to production, on purpose.
Strengths: Translating complexity into decisions, pattern recognition across tech, business, and "why are we doing this?" moments, tireless reviewer and rubber duck
Weaknesses: Not always right, no pager duty, unable to drink coffee despite strong brand alignment
