High performance computing C++ and python engineer position at LPP

Annonce transmise par Roch Smets (LPP)

Le LPP recrute un ingénieur d'étude en CDD (1 an) pour participer au développement d'un code (PHARE) hybride avec maillage adaptatif

Context and environment :

The laboratory of plasma physics (LPP) at CNRS develops codes for simulating plasma processes and reproduce results from laboratory experiments and spacecraft measurements. The developed codes are designed so to be executed on future exa-scale machines and used by world-wide scientific communities. They thus need both high performance and a high level of ergonomy. The LPP is looking for a passionate engineer expert in modern C++ and python development to contribute to one of our main codes.  The development team is small, enjoyable, international, informal and passionate. Your office will be next to other team members in the LPP building in Palaiseau, France.

The mission :

You will join the 3 other devs and participate to the project at all levels from design to implementation, tests etc. The code is open source. It’s about 75% C++ (computation part) and 25% python (user API). The code is parallelized with MPI. On going work is made to also use GPU and multithreading. Depending on your skills and will, you could work on optimizations, parallelization, validation, testing, parallel I/O, user interface, post-processing, etc.

your profile :

Even if it’s better, you do not need a scientific/physics background. But you need to be passionate about writing C++ and python. You worry about the performance and the expressivity of your code (even your C++ templates). You are comfortable with git and the code review / pull request github workflow. You’re autonomous with Linux. Whatever your dev tooling, whether you are Vim or Emacs, QtCreator or VSCode, you master them and can easily navigate through code. There is no particular education background required, only your skills and motivation will be key.

Techniques:

C++17, python, CMake
MPI, SAMRAI, pybind11, highfive/HDF5,
qtcreator, vim
github, teamcity

Duration : 1 year

Salary: CNRS study engineer, public grids, depending on previous experiences (~2k€ net/month)

Contact: nicolas.aunai@lpp.polytechnique.fr