Post-Training Engineer - Apertus — EPFL
- Location
- Lausanne
- Contract
- full-time
- Posted
- 73 days ago
Role overview
Introduction
The Apertus project, a joint effort between EPFL and ETH Zürich, is seeking a practical and motivated engineer to help build the next generation of open foundation models.
The successful candidate will help develop and run post-training and reinforcement learning pipelines for the Apertus project.
- Introduction
- The Apertus project, a joint effort between EPFL and ETH Zürich, is seeking a practical and motivated engineer to help build the next generation of open foundation models.
- Main duties and responsibilities
- The engineer will contribute to the development, execution, and evaluation of scalable post-training workflows for Apertus. Infrastructure and systems engineering
Main responsibilities
- Main duties and responsibilities
- The engineer will contribute to the development, execution, and evaluation of scalable post-training workflows for Apertus. Infrastructure and systems engineering
- Build and maintain containerized 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 utilization.
- Help maintain reproducible training recipes, configuration files, launch scripts, and documentation.
- Work with researchers and CSCS engineers to improve the reliability and performance of large-scale experiments.
- LLM post-training and Reinforcement Learning
- Support SFT, preference optimization, and reinforcement learning workflows.
Additional details
- The engineer will contribute to the development, execution, and evaluation of scalable post-training workflows for Apertus. Infrastructure and systems engineering
- Debug common post-training issues, including optimization instability, reward hacking, regressions, and evaluation failures.
- Strong collaboration and communication skills and ability to work across research and engineering teams. Strongly preferred
- Experience with frameworks such as veRL, slime, Megatron-LM, DeepSpeed, TRL, vLLM, SGLang, or similar tools.
- Experience with large-scale evaluation pipelines.
Notes and original content
- The engineer will contribute to the development, execution, and evaluation of scalable post-training workflows for Apertus.
- Infrastructure and systems engineering
- Strong collaboration and communication skills and ability to work across research and engineering teams.
- Strongly preferred
- Nice to have
Questions about this listing
What salary does EPFL offer for this role?
EPFL lists CHF 60'500 - 91'500 gross per year for this position in Lausanne. This is the salary published in the original listing (or, when the employer omits a figure, a realistic estimate for the role and sector) — the calculator on this site converts it to your actual net take-home once cross-border tax and social contributions are applied.
Is this a full-time role, and what type of contract does EPFL offer?
This listing is a full-time position. The contract type shown here comes directly from the employer's original posting; always confirm exact hours, notice period and probation length with EPFL during the application process, since these details can vary by role even within the same contract category.
Do I need a cross-border work permit for a role in Vaud?
EU/EFTA residents living in the border zone of the country adjoining Canton Vaud can apply for a G permit; the Swiss employer files it with that canton's migration office after the contract is signed. Border-zone rules and processing times vary by neighbouring country and canton, so confirm the specifics with Vaud's cantonal migration office or with HR during the application.
How do I apply for this position at EPFL?
Use the "Apply now" button on this page — it links directly to EPFL's original listing at careers.epfl.ch, so your application goes straight to the employer's own applicant-tracking system. Frontaliere Ticino does not collect or forward applications itself.
