Apertus Engineer: Infrastructure

Join Our Innovative Team

We are looking for a talented individual to enhance our dynamic team. You will take ownership of the container image stack supporting our pre-training, post-training, and serving workloads, while collaborating closely with engineers to ensure large-scale training remains stable and efficient.

About the Project

The Apertus project is a groundbreaking collaboration involving EPFL, ETH Zürich, and CSCS, aimed at developing cutting-edge open foundation models utilizing a sophisticated supercomputing environment. Train models with hundreds of billions of parameters using thousands of GPUs on one of Europe’s largest AI-ready supercomputers. Our dedicated team of engineers works alongside leading researchers and maintains collaborations with over thirty academic institutions to provide responsibly trained, multilingual, and multimodal AI models.

Key Responsibilities

• ML System Image Maintenance: Build, manage, and enhance container images throughout core ML development phases. Ensure reproducibility and thorough documentation, validating images in collaboration with engineers.
• Compute Partnership and Efficiency: Act as the main technical liaison regarding computational efficiency, collaborating with CSCS staff to improve core systems and document best practices for high-performance compute resources.
• Infrastructure Stress Testing: Conduct stress tests on infrastructure, debugging cluster-level issues affecting stability and throughput, and supporting the serving stack.

Essential Qualifications

• MSc or PhD in Computer Science, Data Science, Artificial Intelligence, Machine Learning, or related fields. Exceptional BSc candidates with notable engineering experience will also be considered.
• Hands-on experience with HPC environments, job schedulers, shared filesystems, and multi-node GPU systems.
• Strong Linux systems and container skills, with a focus on Docker/Podman and HPC runtimes.
• Outstanding collaboration and communication skills, able to bridge research and engineering teams.
• Proven experience in relevant domains, either through projects or formal work settings.

Preferred Qualifications

• Familiarity with LLM training and serving frameworks.
• Experience with ARM64/aarch64 platforms and HPC networking stacks.
• Background in parallel filesystems and storage performance optimization.
• Experience in establishing CI/CD pipelines for container image development.

Benefits and Opportunities

• An enriching academic environment at a top-tier technical university.
• Access to one of the largest AI-ready supercomputers in Europe.
• Collaboration with esteemed researchers and engineers from prominent institutions.
• Competitive employment conditions with comprehensive benefits, including pension plans from ETH Zürich/EPFL.
• Flexible working arrangements and professional development opportunities.
• Engagement in open-source projects with significant global impact.
• Contributing to national AI development initiatives.

Explore a unique opportunity to be part of a pioneering project that holds national significance, based in either Lausanne at EPFL or Zürich at ETH Zürich.