Domain Architect - Compute

Domain Architect – AI Compute
12 Month Contract
Sydney or Melbourne
$1,100 - $1,350 p/d

About the opportunity:
A major System Integrator building large scale, purpose built AI infrastructure for enterprise clients is looking for a Domain Architect to take ownership of the compute layer across their NVIDIA based deployments. It's a genuinely hands on role, split roughly 60% delivery and 40% pre sales support, spanning SuperPOD, BasePOD and equivalent AI factory builds. You'll move between low level cluster configuration and client facing conversations in the same week, so both skill sets matter equally.

What you'll be doing:

• Running physical build and provisioning of GPU clusters across NVIDIA's rack scale and server platforms, using cluster management tooling for diskless boot, firmware and OS hardening
• Setting up monitoring and automated provisioning so hardware reaches production state with minimal manual touch
• Configuring GPU scheduling and fair usage policies (fractional GPU allocation, preemption, multi tenant quotas) across schedulers like Slurm, Run:AI, Kueue and Volcano
• Making sure workload placement respects underlying hardware topology for consistent performance
• Standing up multi cluster observability and GPU health telemetry to catch thermal and error issues early
• Validating cluster performance with industry standard HPC benchmarking tools, and tuning the OS and kernel for high throughput fabric performance
• Owning technical accuracy of the compute components list, and supporting the sales team with scoping, estimation and discovery workshops for upcoming deployments

What we need to see:

• Strong architectural knowledge of current and recent generation NVIDIA GPU platforms, including rack scale and NVSwitch fabric integration
• Expert level Linux engineering across Ubuntu and RHEL in HPC or AI contexts, with real kernel tuning and hardening experience
• Solid scripting and automation background in Python and Ansible for infrastructure configuration
• Comfortable owning technical scope and BoM accuracy, and confident running client workshops as well as doing the build work

Bonus points:

• Prior experience inside a System Integrator or Managed Service Provider
• Hands on time with cluster management platforms similar to Bright Cluster Manager / Base Command Manager
• Understanding of high speed interconnect fabrics and how they interact with host level PCIe and NVLink topology
• Exposure to liquid cooling infrastructure and its maintenance implications
• Kubernetes or OpenShift administration experience
• Public cloud compute design experience for AI workloads (Azure, AWS, GCP, OCI)
• Familiarity with multi cluster management tooling, AI workload orchestration platforms, or digital twin/rendering environments