AI Infrastructure Engineer

About the role

The AI Infrastructure Engineer will design, build and maintain high‑density GPU compute clusters that power AI/ML workloads across the organisation. You will work closely with AI/ML teams to ensure optimal performance, reliability and scalability of the underlying infrastructure.

Key responsibilities

• Architect, configure and operate multi‑GPU compute clusters (e.g., NVIDIA HGX/DGX) and associated container orchestration platforms such as Kubernetes, Slurm or Ray.
• Monitor GPU health, telemetry and utilization, and implement strategies to minimise idle time and bottlenecks.
• Design and optimise low‑latency network fabrics (InfiniBand, RoCE v2, NVLink) and high‑throughput storage solutions (Lustre, GPFS, Ceph, MinIO, NVMe‑oF) for distributed training.
• Develop automated deployment pipelines using Infrastructure as Code tools (Terraform, Ansible, Helm, Pulumi) and maintain golden images and firmware updates.
• Implement end‑to‑end observability with Prometheus, Grafana, DCGM exporter and NVIDIA System Management Interface, and troubleshoot performance issues with AI/ML teams.

Required profile

• Bachelor’s degree in Computer Science, IT, Computer Engineering or equivalent practical experience.
• 3–6+ years of hands‑on experience in infrastructure engineering, high‑performance computing, DevOps or cloud infrastructure.
• Relevant certifications (e.g., CKA/CKAD, NVIDIA Certified Associate/Professional, AWS/Azure/GCP Solutions Architect) are a plus.

Required skills

• Linux systems administration, kernel tuning, Bash and Python scripting.
• GPU hardware architectures, CUDA, cuDNN, PCIe/NVLink.
• Kubernetes (GPU operator, device plugins), Slurm, Run:ai, Ray.
• RDMA, RoCE v2, InfiniBand, Priority Flow Control (PFC), ECN.
• High‑IOPS storage systems: Lustre, GPFS/IBM Spectrum Scale, Ceph, MinIO, NVMe‑oF.
• Infrastructure as Code: Terraform, Ansible, Helm, Pulumi.
• Monitoring and observability: Prometheus, Grafana, DCGM exporter, NVIDIA SMI.