Infrastructure / DevOps Engineer

About the role

We are seeking an Infrastructure / DevOps Engineer to join a leading 3D scanning company that combines cutting‑edge computer vision and machine learning to produce high‑quality, relightable 3D models at scale. You will be responsible for building and maintaining the core compute infrastructure that powers multi‑GPU clusters across cloud and on‑premise locations.

Key responsibilities

• Provision, configure, and maintain heterogeneous CPU/GPU compute clusters in multiple physical sites.
• Design and implement dynamic compute and storage provisioning based on workload demands.
• Develop storage solutions (NAS, distributed filesystems, tiering) at hardware and software levels.
• Implement and manage container orchestration platforms such as Kubernetes and Docker.
• Write infrastructure‑as‑code using Terraform and Ansible.
• Build and optimise job scheduling systems (e.g., Slurm, Kubernetes).
• Set up monitoring, alerting, and observability with Prometheus and Grafana.
• Profile and optimise system performance: GPU utilisation, memory bandwidth, I/O throughput, network latency.
• Manage networking, VPNs, and secure access across distributed environments.
• Handle reliability concerns, including hardware failure detection, job checkpointing, and disaster recovery.

Required profile

• Strong Linux system administration experience.
• Hands‑on experience with Docker and Kubernetes.
• Proficiency with infrastructure‑as‑code tools such as Terraform and Ansible.
• Experience managing HPC clusters and job schedulers like Slurm.
• Familiarity with monitoring tools (Prometheus, Grafana).
• Solid understanding of networking principles and secure remote access.
• Experience with hardware management interfaces (IPMI, BMC) and server maintenance.
• Knowledge of storage system design (NFS, Ceph, distributed filesystems).

Required skills

• Linux
• Docker
• Kubernetes
• Terraform
• Ansible
• Slurm
• Prometheus
• Grafana
• IPMI
• BMC
• NFS
• Ceph
• AWS (optional)
• MaaS (bare‑metal provisioning, optional)