CloudEngine Digital Hiring! Full Time AIDC Infra Delivery Engineer (Network - Server) in Federal Territory - Ricebowl

Job Overview

As the on-site technical lead for overseas projects, you are responsible for full delivery and acceptance of NVIDIA GPU clusters including H100, H200, GB200, GB300, B300 etc. You follow NVIDIA NVIS delivery standards, and take charge of project technology, progress, quality and customer satisfaction.

Core Responsibilities

1. Act as exclusive technical SPOC for overseas projects, control overall technical solutions, risks, milestones and acceptance standards.

2. Complete site survey, rack layout design, network topology planning and detailed cabling scheme formulation.

3. Carry out GPU server installation, RoCEv2 / InfiniBand network deployment, system, driver and related software configuration.

4. Conduct performance benchmark testing, finish project verification and deliver acceptance reports.

5. Troubleshoot GPU and network complex faults, sort out RCA solutions.

6. Provide English technical documents, operation guidelines and professional training for clients.

7. Coordinate internal and external resources to push project issues solved efficiently.

Requirements

1. 3-5 years experience in AI HPC and data center field; finished at least 3 complete GPU cluster delivery projects.

2. Master NVIDIA GPU hardware architecture, proficient in RoCE & IB network deployment and tuning.

3. Skilled in Linux, CUDA, NCCL, DCGM, OFED and mainstream AI cluster scheduling tools.

4. Have overseas project delivery experience, fluent in English for technical communication and meeting presentation.

5. Strong stress resistance and project execution, able to adapt cross-cultural working environment.

6. Available for overseas business trips to Singapore, Thailand, Indonesia, Japan, US and other countries.

Preferred

• Own NVIDIA official certification

• Experienced in liquid cooling GPU cluster deployment

• Have working experience with global cloud vendors and large-scale AI computing projects