Major duties and responsibilities:
- Bridge the gap between the HPC systems and the researchers.
- Participate to HPC coursesand outreach.
- Support the preparation of production project proposals, such as national computing allocation projects at CSCS or DOE Incite.
- Help users in the migration from workstation to compute cluster and from compute cluster to high-end infrastructure.
- Help optimize code: sequential efficiency, parallelization, scaling, and algorithm choice.
- Enable effective usage of the pooled computing resources by collaborating to the installation of specialized software.
- Support the EPFL HPC users community with application-related expertise.
- Evaluation and benchmarking of new hardware technology.
- HPC technology survey.
Essential skills and experience required:
- PhD in science or engineering.
- Extensive experience (at least 6 years) in optimization of scientific software and parallel programming (OpenMP, MPI,…).
- Experience in debugging parallel applications at scale.
- Extensive experience in teaching HPC related topics.
- An avid interest in science and engineering applications for academia and new computing paradigms.
- Proven ability to solve problems independently.
- Good team player.
- Proficiency in both English and French.
What we offer:
- A young, dynamic, inter-disciplinary, challenging academic and multicultural environment.
- An opportunity to evolve in a framework promoting advanced technology.
- Excellent working conditions and competitive remuneration.
- Opportunities for continuing education.
Start date: 01.06.2014 (or earlier)
Deadline for application: 15.04.2014
Activity rate: 100%.
Duration of contract: permanent and non permanent (2 years)
Contact: Further information can be obtained by contacting Vittoria Rezzonico, SCITAS Executive Director, by E-mail: Vittoria.Rezzonico@epfl.ch.
Applicants should submit a cover letter and a detailed CV in PDF format only, electronically to email@example.com. Please use the position title and the reference in the “subject” field.