Platform Engineer / DevOps

The team operates a production AI platform running large language model workloads for the technology industry. The platform uses Kubernetes on Talos Linux with GPU-enabled nodes and technologies such as NVIDIA NIM and vLLM.

• Strong hands-on experience operating Kubernetes in production environments.
• Deep practical understanding of Kubernetes infrastructure and troubleshooting.
• Experience with cluster lifecycle management and production upgrades.
• Strong Linux administration skills.
• Experience owning production systems and resolving complex incidents.
• Understanding of Kubernetes Operators, controllers, and CRDs.
• Strong troubleshooting and root cause analysis capabilities.
• Good understanding of observability, monitoring, and production reliability.
• Ability to work independently and take technical ownership.
• Fluent spoken and written English.

Nice to have:

• Experience with Talos Linux.
• Experience running Kubernetes on bare metal or self-managed infrastructure.
• Experience with GPU-enabled Kubernetes environments.
• Knowledge of NVIDIA GPU Operator, NVIDIA NIM, CUDA, or similar NVIDIA technologies.
• Experience with vLLM or other model-serving / inference technologies.
• Background in AI/ML infrastructure, MLOps, or AI platforms.
• Swedish language skills or willingness to develop Swedish over time.

• Operate and continuously improve a production-grade Kubernetes platform.
• Take end-to-end ownership of selected platform components.
• Troubleshoot complex Kubernetes infrastructure and production issues.
• Investigate and resolve issues such as unavailable nodes, unhealthy or hanging pods, Kubernetes Operators, CRDs and controllers, unexpected reconciliation behaviour, platform and infrastructure dependencies.
• Restore production services and perform root cause analysis.
• Implement improvements to prevent recurring incidents.
• Plan and perform safe rolling upgrades of Kubernetes clusters and platform components.
• Support highly available and predictable operation of GPU-enabled infrastructure.
• Operate infrastructure supporting LLM inference and model serving using technologies like NVIDIA NIM and vLLM.
• Improve automation, observability, and operational processes.
• Create reusable platform patterns and clear interfaces for product teams.
• Improve developer experience and reduce time-to-market.
• Maintain strong security and data residency standards.
• Collaborate closely with engineers, infrastructure teams, technical leads, and business stakeholders.