Senior Platform Engineer

About the project

A European AI company is building a platform for running AI models. It has to install wherever its customers need it, in public cloud, private and sovereign cloud, on-prem Kubernetes and networks with no internet access. We are looking for an engineer to own how the platform is packaged, installed, upgraded and connected to customer infrastructure.

What you will do Build and maintain the Helm charts, operators and offline bundles that install and upgrade the platform across Kubernetes distributions and clouds, including air-gapped networks. Connect the platform to customer infrastructure, including OIDC and SAML identity, object storage, container registries, ingress and network policy, and GPU scheduling. Write the Python services and REST APIs through which third-party open-source apps are packaged, installed and managed inside the platform. Own CI/CD, releases and the test matrix across supported environments, add metrics, logs and traces, and write the runbooks.

Must have 5+ years in platform, infrastructure or backend engineering, 2+ of them running production workloads on Kubernetes. 3+ years of production Python backend work with FastAPI or Django REST Framework, Pydantic, SQLAlchemy and Alembic on PostgreSQL, async, pytest and type hints. Helm charts that other people install, written and maintained by you. Installs into environments you do not control, such as customer on-prem, private cloud or air-gapped networks. AWS (EKS, ECR, RDS, S3, Secrets Manager), OIDC, OAuth2 or SAML, Docker image hardening, GitLab CI or GitHub Actions. You live in an EU country and can work on a B2B contract. English at B2+ for daily calls with the client team.

Nice to have Kubernetes operators (kopf or Go), ArgoCD or Flux. GPU on Kubernetes (NVIDIA GPU Operator), vLLM or Triton. Prometheus, Grafana and OpenTelemetry. Published Python packages or CLIs, contributions to open-source infrastructure.

Terms Full-time and long-term, remote from the EU, B2B contract with New Wave Devs, paid monthly. Occasional trips to the client's office in Europe are possible. We are shortlisting now for a client project, and the start date depends on the client.

How we hire First, the three screening questions below. Then a 30 to 45 minute technical call with our CTO, screen shared, without AI assistants. After that, an interview with the client team.