HPC Performance Engineer,Frontier,NSCC

A*STAR - Agency for Science, Technology and Research; Singapore; On-site; SGD 90,000 - 150,000; Full time

Job summary

A*STAR NSCC in Singapore seeks an experienced HPC Performance Engineer to optimize the new national supercomputer for peak performance and energy efficiency across AI, climate and materials science domains.

You will profile user applications, develop scalable HPC software, and collaborate with researchers to influence future NSCC architectures. Strong C/C++, Python, MPI/OpenMP and GPU programming are essential.

Qualifications

• Bachelor's degree in computer science, computer engineering, or related field.
• 3 years of experience in parallel programming or performance optimization.
• Experience with HPC software stacks and development toolchains; Spack or EasyBuild is a plus.
• Strong C/C++ and Python programming; Fortran a plus.
• Familiar with OpenMP/MPI, Linux, HPC profilers and debuggers.
• Familiar with GPU programming.
• Familiar with container technologies.

Responsibilities

• Conduct performance engineering for large-scale HPC systems and workloads.
• Profile and model performance of user applications and scientific software stacks.
• Design, develop and implement HPC software best practices and scalable software infrastructure.
• Provide HPC and scientific domain advice to NSCC users.
• Collaborate with researchers across AI, climate, and materials science.
• Assist in planning and designing future HPC systems and benchmarking NSCC workloads.
• Analyze system and user job data for efficient resource allocation and management.
• Develop HPC utilities and automated testing tools.
• Create HPC user guides and best practice documents.
• Stay up-to-date with scientific domain research and HPC technologies.

Skills

C/C++

Python

GPU programming

MPI/OpenMP

Linux

HPC profiling

Containers

Spack/EasyBuild

Education

Bachelor's degree in computer science or computer engineering

Tools

GPFS/Lustre

Job description

About The Role

Join the team responsible for Singapore's new supercomputer. As our HPC Performance Engineer, you are the "engine tuner" for the entire national facility. Your work will ensure that this multi-million-dollar investment runs at peak performance and efficiency for all of Singapore's key research domains from AI to climate and material science. This is a unique opportunity to tackle diverse performance challenges at a massive scale, profile applications of national importance, and directly influence the design of future national HPC systems.

Responsibilities

• Conduct performance engineering to large‑scale HPC system and application workloads, optimising both the performance and energy efficiency.
• Profile and build performance models of the user applications, workflow and scientific software stacks.
• Design, develop and implement HPC software best practices and scalable software infrastructure.
• Provide HPC and scientific domain advice to users of NSCC systems.
• Engage and collaborate with new researchers, communities, and disciplines with computationally intensive requirements.
• Assist in the planning and design of future HPC systems, including benchmarking NSCC workloads on various platforms and recommending the most suitable architecture for the research community.
• Analyse system and user job data for efficient resource allocation and management.
• Develop HPC utilities and automated testing tools for NSCC HPC systems.
• Develop HPC user and best practice guides for NSCC HPC systems.
• Get up‑to‑date with scientific domain research development, HPC system and software technology.

Qualifications

• Bachelor's degree in the field of computer science, computer engineering, or other relevant areas.
• Ideally, 3 years of experience in parallel programming or performance optimisation.
• Experience with development toolchain and scientific application software stack for building HPC software is essential; knowledge in using Spack or EasyBuild is a plus.
• Strong programming skills in C/C++ and Python; familiar with Fortran programming is a plus.
• Familiar with parallel and/or GPU programming is highly desired.
• Familiar with Linux environment, scripting languages, HPC profiler and debugger tools.
• Familiar with HPC job schedulers, OpenMP & MPI, and container technologies.
• Familiar with HPC storage, e.g. GPFS and Lustre.
• Familiar with the design & development of multi‑core CPU and GPU.
• Demonstrated team player with strong problem‑solving skills.
• Demonstrated effective communication skills including the ability to articulate technical concepts to a diverse range of audiences.
• Demonstrated ability and willingness to contribute novel ideas and approaches in support of the research community.
• Demonstrated passion for continuous learning and exploring new technologies or domains.