DevOps Engineer

SWYCS builds a smart-building platform: energy monitoring, room reservations and IoT device management for offices and commercial real estate. Thousands of sensors and gateways talk to our platform over MQTT and LoRaWAN.

Our previous DevOps engineer has left and we want the work off our developers’ plates. The infrastructure runs and is documented. We need someone whose job it is to keep it that way.

What you’ll be doing

This is a maintenance role, not a greenfield project.

• Keep our servers healthy. We run on European VPS providers (TransIP, Hetzner): two Kubernetes clusters and around twenty VMs. Some older hosts still need OS upgrades, you plan and execute those.
• Do upgrades on a regular cadence: Kubernetes versions, Helm charts, Postgres, ClickHouse, Redis and the rest of our self-hosted services. Small and boring beats overdue and scary.
• Run vulnerability scans, triage the results and patch what matters.
• Watch disk space, certificates, backups and queues. Be the first responder when a runner fills up or a broker misbehaves. Write a short post-mortem when something breaks.
• Keep backups running and prove a restore works every now and then.
• Support the developers. They deploy their own code through GitLab CI. You keep the runners, registries, clusters and databases they depend on reliable.

There is a roadmap with improvement work for when you have room, but the core of the job is that everything keeps running without a developer having to think about it.

Our stack

• Kubernetes (TransIP managed and self-managed kubeadm), Docker Compose on VMs, Envoy Gateway, cert-manager
• ArgoCD, Helm, SOPS, GitLab CI with self-hosted runners. Legacy production is Salt-managed.
• Databases and middleware, all self-hosted: PostgreSQL (CloudNativePG, pgbouncer), ClickHouse, Redis, NSQ, MinIO, Meilisearch
• IoT: MQTT brokers (Mosquitto, VerneMQ), ChirpStack/LoRaWAN, an on-prem gateway fleet
• Monitoring: Sentry, Prometheus, Loki

• You have kept self-hosted infrastructure alive before: Linux servers, Kubernetes and databases, without a managed cloud to lean on.
• You can work in a mixed estate. Half is modern (GitOps, managed Kubernetes), half is legacy (Compose VMs, Salt, older Ubuntu). You don’t need the legacy half rewritten before you can maintain it.
• Solid bash, and enough scripting to automate chores you find yourself repeating.
• You work in a structured way. You log what you changed and why, and upgrades happen on a schedule, not when things catch fire.
• You handle autonomy well. With 20 hours a week you set your own priorities within a clear mandate.

Nice to have

• SaltStack, ClickHouse operations, MQTT/LoRaWAN, Postgres failover experience
• Experience as the only ops person for a small development team

20 hours per week, schedule largely up to you. Some availability for incidents matters more than fixed hours.

Interested? Tell us briefly how you would take over a running infrastructure from someone else. That tells us more than a list of certifications.

Salaris: €3.500,00 - €4.500,00 per maand