HPC Engineer (Compute)

Desirable Experience

• Networking: Experience with high-speed interconnects (InfiniBand NDR/HDR, RoCEv2) and debugging connectivity issues.
• Containerisation: Experience with Docker and Kubernetes (specifically the NVIDIA GPU Operator).
• Cisco Integration: Familiarity with Cisco UCS or Cisco Nexus configurations in an AI context.

Certifications

Highly Desirable:

• NVIDIA-Certified Associate: AI Infrastructure and Operations (NCA-AIIO)
• NVIDIA-Certified Professional: AI Infrastructure (NCP-AII)
• Red Hat Certified System Administrator (RHCSA)

Success Metrics (KPIs)

• Deployment Velocity: Achieving <24 hour turnaround from "Rack Handover" to "OS Provisioned" for assigned nodes.
• Validation Accuracy: 100% of handed-over nodes pass the "Golden Config" validation script (zero "Dead-on-Arrival" nodes handed to the client).
• Ticket Efficiency: Consistently meeting SLAs for L2 support tickets.

#LI-RB1

About the Role

The HPC Engineer - Compute acts as the primary execution engine for the physical and logical lifecycle of high-performance GPU compute fleets. While the Domain Architect designs the "Gold Standard" and the Squad Lead directs the workflow, you are the "Builder" responsible for the hands-on configuration and validation of the infrastructure. You are a "doer" who is as comfortable flashing firmware on an NVIDIA B300 in the CLI as you are debugging a failed Ansible playbook.

As a System Integrator, we do not simply manage a static cloud; we design and deliver bespoke, high-scale AI factories for the world's leading enterprises. In this role, you will move beyond standard server administration to execute the repeatable, scalable, and automated deployment of NVIDIA SuperPOD, NVIDIA BasePOD, and Cisco AI Factory environments.

In this role, you will operate with a 100% focus on Delivery, executing the Low-Level Designs (LLD) assigned by your Squad Lead. You will own the "Build" phase in the critical "Compute-Network-Storage" triad, ensuring our clients receive defect-free, validated infrastructure.

Key Responsibilities

1. Build & Provisioning (Execution)

• Cluster Assembly: Execute the automated provisioning of bare-metal AI nodes (DGX H100/Blackwell, HGX, IGX, MGX, NVL72) using NVIDIA Base Command Manager (BCM) (formerly Bright Cluster Manager).
• Infrastructure as Code (IaC) Execution: Run and maintain the Ansible playbooks required to configure Operating Systems (RHEL/Ubuntu), inject user authentication (LDAP/AD), and mount high-performance parallel file systems.
• Hardware Lifecycle: Perform complex firmware upgrades (SBIOS, BMC, GPU VBIOS, NVSwitch) across massive clusters without disrupting operations, ensuring strict adherence to the NVIDIA Firmware recipe.
• Network Integration: Configure compute-side networking (IPoIB, Netplan) to align with the Ethernet or InfiniBand fabric specifications provided by the Network Domain.

1. Validation & Testing

• Performance Benchmarking: Execute and log standard acceptance tests, including HPL (High-Performance Linpack) and NCCL-tests, to verify node performance against the "Gold Standard" benchmarks.
• Health Checks: Run "Burn-in" scripts on new hardware to identify "Dead on Arrival" components (e.g., faulty GPU memory, loose NVLink cables, Xid errors) before client handover.
• Test Reporting: Generate "As-Built" validation reports that the Field Solutions Engineers (FSEs) use for final client sign-off.

1. Operations & Support

• Ticket Resolution: Handle L2 support tickets escalated from L1, such as "Stuck Slurm Jobs," "GPU falling off the bus," "ECC Errors," or "Node not draining."
• Routine Maintenance: Execute scheduled maintenance windows for kernel patching, driver updates, and security hardening.

Technical Competencies

Essential Skills

HPC & AI Infrastructure:

• Cluster Provisioning: Proficiency with NVIDIA Base Command Manager (BCM) for bare-metal provisioning and image management.
• Hardware Architecture: Deep understanding of NVIDIA DGX/HGX/NVL72/MGX architectures, including PCIe topology, NVLink/NVSwitch connectivity, and GPU memory hierarchy.
• Workload Management: ability to troubleshoot basic Slurm issues (job dependencies, partition misconfigurations, node draining).

Linux & Automation:

• Linux Administration: Solid mastery of RHEL/Ubuntu internals, including systemd, kernel modules, and package management.
• Automation: Ability to read and execute Ansible playbooks and write basic Python/Bash scripts for task automation.
• Version Control: Familiarity with Git workflows (pulling code, creating branches, committing config changes).