HPC Applications Specialist

Do you want to move science forward? Solve complex problems and turn them into simple and elegant solutions? Work on something meaningful with an innovative team that creates value for our society?

Then ETH Zurich is looking for you.

Our mission in the High Performance Computing group is to provide first-class HPC and AI/ML resources to researchers from all departments of ETH Zurich. We operate compute clusters with thousands of CPUs and GPUs, petabytes of storage, high-speed / low-latency networks, in multiple data centres in Zurich and Lugano. Since our users' needs are constantly evolving, we are using cutting-edge, cloud-native technologies like Kubernetes to expand our services to world-class supercomputers and into the cloud.

We are looking for an Applications specialist with a passion for HPC.

This position is part of our "NextGen HPC" project, which will enable ETH Zurich to expand its "Euler" computing platform to various HPC infrastructures in Switzerland and/or abroad. As part of this project, your role will be to port a series of scientific applications to new hybrid CPU+GPU architectures like Nvidia Grace Hopper and AMD Instinct APUs. This includes compilation, testing, debugging, performance analysis (efficiency, scaling), and optimization. You will make these applications available to ETH researchers using Spack and Singularity/Apptainer, and automate the installation process using industry-standard CI/CD pipelines based on Kubernetes. You will also assist ETH researchers in the form of tutorials, workshops and personalized user support.

These tasks will require a close collaboration with the DevOps and User Support teams in the HPC group.

Requirements:

• BSc. Computer Science or related
• Ability to understand scientists’ computing needs and to address them
• First-hand experience with large-scale scientific computing applications
• Experience with parallelization and optimization of scientific codes
• Compilation and packaging of complex open-source codes using e.g. Spack and Singularity
• Coding experience with Bash and Python
• Familiarity with C/C++, Fortran, MPI and openMP
• Ability to learn new things quickly and put them into practice
• Good communication skills in English

The following are considered an advantage:

• Programming experience with CUDA, ROCm, PyTorch, TensorFlow
• Operation of Linux servers and large-scale clusters
• Understanding of Kubernetes and cloud-native technologies
• Experience with hybrid CPU+GPU systems (Nvidia Grace-Hopper, AMD Instinct or equivalent)

• Dynamic international team
• Brand new offices in Zurich with free coffee
• Hybrid working model (home / office)
• Half-price access to public transport in Switzerland
• Possibility to attend ETH lectures and courses
• Participation to conferences in Switzerland and abroad

This position is limited to one year. An extension may be possible subject to budgetary considerations.