Systems Engineer (AI/Linux) — Server Deployment - #1118561

Location: Singapore
Employment: Full‑time
Reports to: IT Support Manager

Role Summary

We deploy high‑performance AI servers for enterprise and public‑sector clients. You’ll assist the IT Support Manager to stage, image, validate and install single‑node and small‑pod systems. The role is hands‑on in the lab and onsite at data centres, with a strong focus on consistency, documentation and speed to customer acceptance.

Key Responsibilities

• Staging & Rack‑ready Prep: Rack, cable, label; capture asset/serial info and maintain inventory sheets.

• OS & Driver Stack: Install Ubuntu/Rocky Linux, NVIDIA drivers, CUDA/cuDNN; verify with nvidia-smi.

• Firmware & BIOS: Apply vendor‑approved BIOS/BMC/NIC firmware baselines; keep a simple firmware bill per build.

• Imaging & Automation: Run PXE/imaging flows and Ansible/bash scripts; follow checklists to prevent drift.

• Burn‑in & Validation: Execute 24–48h burn‑in; run sanity tests (NCCL ring‑allreduce, fio/iperf, stress/thermals) and record results.

• Onsite Deployment: Assist with rack‑in, power/network checks, SAT/UAT, and collect same‑day customer sign‑off.

• Documentation: Produce acceptance packs (serials, firmware bill, test logs, photos); update runbooks and templates.

• Spares & RMA: Prep spares kits, coordinate RMAs and returns with vendors and logistics.

• HSSE & DC etiquette: Follow ESD, lifting and DC safety rules; maintain tidy work areas and cabling standards.

Requirements

• 1–3 years in Linux/servers or DC operations; poly/degree or equivalent experience.

• Comfortable with Linux CLI, systemd, basic networking (VLAN/IP), and SSH.

• Familiar with NVIDIA stack (drivers, CUDA, basic NCCL checks) and firmware updates (BMC/BIOS/NIC).

• Able to follow and improve checklists and scripts (bash/Ansible basics).

• Physically able for DC work (rack equipment, cabling); clean documentation habits.

• Good communication; customer‑facing onsite when required.

Nice‑to‑Haves

• Experience with Supermicro platforms and Mellanox/ConnectX NICs.

• PXE/kickstart/cloud‑init experience; Grafana/Prometheus for quick burn‑in dashboards.

• Basic Windows Server installs for mixed environments.