The posting
Opportunity
Hybrid (2 days on-site) · Permanent, full-time
About the role
The data stack you own runs on warships and in air-gapped ops centres. When ILIAS Solutions cannot be in the room, your Helm charts, runbooks and Spark jobs are what keeps it running, across 10+ customer sites, some physically unreachable, where what you deploy must continue to function for years.
This is not a build-and-hand-off role. You will sustain, evolve and set the engineering standard for a data runtime covering Spark on Kubernetes, Delta Lake on MinIO, and PostgreSQL, feeding dashboards, analytical notebooks and a GenAI assistant, while ensuring every component can be installed and operated by customer staff without your presence.
Your Mission
What you'll do
- Own the full data runtime in production: deploy, configure, monitor and upgrade Spark (Operator or spark-submit), Delta Lake on MinIO, and PostgreSQL on Kubernetes, including backup and restore, health monitoring and executor-level diagnosis across sites you may never be able to log into.
- Build and operate Spark processing jobs end-to-end with Analytics Engineers, from schema design through driver and executor-level diagnosis, and maintain Delta Lake hygiene: schema evolution, compaction, vacuum and small-file control.
- Design and steward PostgreSQL for analytical and serving workloads: schema design, performance tuning, migration strategy, autovacuum and memory configuration, and tested restores using pgBackRest or Barman.
- Deliver backend services in Python and Java that feed dashboards, analytical notebooks and the GenAI assistant, with integration tests against real PostgreSQL and MinIO and reproducible local environments.
- Package and publish the full data stack as Helm charts and signed offline bundles, installable by customer staff without ILIAS present, and lead the migration of data services from Docker Swarm to RKE2.
- Raise the chapter's capability floor: write runbooks and health checks that function on unreachable sites, set the standard for stateful service deployment and observability, and mentor engineers without becoming a single point of failure.
What success looks like
The data runtime runs reliably at customer sites with no dependency on your physical presence, runbooks are clear enough for customer staff to act, and health checks surface issues before they become incidents. The Docker Swarm to RKE2 migration is complete, tested and documented, and the offline bundle process is reproducible and signed. As your ownership grows, the chapter has a shared standard for stateful service deployment and observability that outlasts any single engineer.
What you bring
- Production operation of Spark on Kubernetes: Operator or spark-submit, driver and executor pod templates, S3A and Delta Lake jar configuration, dynamic allocation, executor sizing, shuffle and spill management, MinIO committers, and Prometheus/JMX monitoring; this is the primary filter for this role.
- Operator-managed PostgreSQL with DBA-adjacent depth: CloudNativePG, Zalando or Crunchy; autovacuum and memory tuning; TimescaleDB and PostGIS; pgBackRest or Barman restores; postgres_exporter monitoring.
- Production operation of MinIO in distributed mode: erasure coding, bucket policies, OIDC via Keycloak, and Prometheus metrics.
- Kubernetes at the application layer: operators, Helm, GitOps (Argo CD or Flux), offline packaging, storage (Longhorn or Rook), NetworkPolicies and RBAC; cluster administration is not required.
- 6+ years of backend engineering in production, 3+ on data-intensive systems; Python at senior level, SQL at expert level, Java at working level; Git at expert level.
- Eligibility to obtain and maintain the required national or NATO security clearance for this role (we support candidates through the clearance process).
If you meet most of the requirements above, we encourage you to apply. The security clearance eligibility requirement is not flexible.
Nice to have
The following experience is particularly valuable given the live production environment; any two of the items below move a candidate to the top of the list:
- Kafka producers and consumers with schema management; Airflow or comparable orchestration; Scala sufficient to read Spark source and stack traces.
- Air-gapped delivery with Zarf or signed offline bundles; STIG-hardened hosts and FIPS-mode library knowledge; Docker Swarm migration experience; remote shuffle services and shuffle-heavy job tuning.
Also valued, and not required:
- JupyterHub, MLflow or vLLM model serving; Go for operators or CLI tooling; Ansible or Terraform; Iceberg or Hudi lakehouse formats.
- Experience in defence or another regulated industry; professional working proficiency in Dutch or French.
Build Your Future with ILIAS
Why ILIAS Solutions
ILIAS Solutions is part of the Patria Group and builds software that defence organisations use to sustain operational readiness. The constraints are hard: air-gapped environments, long-lived deployments, hardware you cannot always reach. The engineering has to be right because the alternative is a system that fails without anyone present to fix it.
You will work in a Backend Engineers chapter within a product engineering tribe, with ownership that extends from design through production operation, and the expectation that the chapter's standards improve durably as a result of your presence.
Practicalities
Location: ILIAS Solutions, Diegem, Belgium. Permanent, full-time position.
*Actual compensation will be determined based on factors such as skills, experience, qualification and internal equity and can ultimately be more or less than the posted salary range.*
Eligibility
Because of the nature of our work, this role may require eligibility under applicable export-control regulations and the ability to obtain national or NATO security clearance. We assess eligibility lawfully and on a role-by-role basis; specific requirements are discussed during the selection process. We do not exclude candidates on the basis of nationality except where mandated by applicable export-control or security-clearance law.
How to apply
Send your application through the vacancy page. Questions before you apply? Reach out to Tim van Dam at [email protected], happy to tell you more about the role and the process, including what the clearance process involves in practice.
