Remote Post-Training LLM Engineer: HPC & RL Pipelines

Immigration Policy Lab; Zürich; Vor Ort; CHF 120.000 - 180.000; Vollzeit

Benefits dieser Stelle

Remote work options

Professional development

Open-source collaboration

Academic environment

Zusammenfassung

ETH Zurich is seeking an Apertus Engineer for post-training to develop, run, and evaluate scalable post-training workflows for Apertus base models, turning them into capable assistants.

You will build containerised environments for LLM post-training and RL workloads, adapt for Alps/CSCS, run Slurm-based training, and collaborate with researchers and CSCS engineers. The role can be based in Zürich or Lausanne, with remote options as applicable.

Qualifikationen

• MSc or PhD in Computer Science, Data Science, AI, ML, or related field.
• Exceptional BSc candidates with strong engineering experience will be considered.
• Experience in AI and neural network architectures.
• Strong collaboration and communication across research and engineering teams.
• Hands-on experience with LLM post-training (SFT, preference optimisation) or reinforcement learning.
• Experience with frameworks such as veRL, slime, Megatron-LM, DeepSpeed, TRL, vLLM, SGLang, or similar tools.

Aufgaben

• Develop, execute, and evaluate scalable post-training workflows for Apertus.
• Build and maintain containerised environments for LLM post-training and RL workloads.
• Adapt containers and dependencies for execution on Alps / CSCS infrastructure.
• Run and monitor Slurm-based training and evaluation jobs.
• Debug failures related to distributed execution, checkpointing, filesystem performance, networking, and GPU utilisation.
• Help maintain reproducible training recipes, configuration files, launch scripts, and documentation.
• Work with researchers and CSCS engineers to improve reliability and performance of large-scale experiments.
• Evaluate model behaviour across reasoning, mathematics, instruction-following, multilingual, tool-use, and safety benchmarks.

Kenntnisse

LLM post-training

Reinforcement learning

HPC environments

Containerization

Python

Distributed training

GPU clusters

Ausbildung

MSc or PhD in Computer Science / AI / ML

Strong engineering background

Tools

Slurm

Megatron-LM

DeepSpeed

veRL

TRL

vLLM

Jobbeschreibung

ETH Zurich is seeking an Apertus Engineer for post-training to develop, run, and evaluate scalable post-training workflows for Apertus base models, turning them into capable assistants.

You will build containerised environments for LLM post-training and RL workloads, adapt for Alps/CSCS, run Slurm-based training, and collaborate with researchers and CSCS engineers. The role can be based in Zürich or Lausanne, with remote options as applicable.