1 New internal mobility in IT-SC
Software Developer in the SC group, RD (Research and Development) section with a background in the use of accelerators (GPUs/FPGAs), working in close collaboration with the LHC experiments and other teams on solving the HL-LHC challenge.
- Software development for the cross-community computing projects
- Research and development of the effective use of computing accelerators
- Development of concepts to improve the portability and sustainability of HEP code
- Contribution to training and dissemination activities in these areas
The main tasks will be:
- Contribute to cross-community development projects. In these projects the team collaborates with the core development teams of the large scientific software packages to improve their performance and adapt them to the changes in the computing infrastructure
- Evaluate and apply libraries and frameworks for the abstraction of hardware accelerators. The basis of these studies will be the work with components of the HEP software stack such as GEANT or event generator code.
- Help the community in the process to make efficient use of modern HPC systems. The focus will be to develop strategies based on practical experience.
- HEP code has a very long lifetime and the code contains contributions of hundred scientists, with many of them being no longer part of the community. Based on your work with these codes you will suggest approaches and methodologies to improve portability and sustainability.
- Prepare training and information material related to your field of expertise and actively communicate the results of your work.
- Actively build collaborative relationships with the developer and user communities of HEP software at CERN, by participating in workshops, conferences and informal exchange of information
Required skills include:
- Software development
- In-depth experience with modern C++ in the Linux/UNIX environment
- Significant development of code that uses accelerators, either GPUs or FPGAs
- Experience in refactoring/porting of a large code base (>100k sloc)
- Some familiarity with components of the HEP software stack (or experience in supporting scientific software in an HPC environment)
- Experience with accelerator abstraction frameworks like Alpaka, Kokos, oneApi, SYCL, OpenCL, OpenMP…
We are looking for a person who can build, or already has, strong relationships with software developers throughout the LHC community.
Being a “Technology Evangelist” would be an asset ;-) (https://en.wikipedia.org/wiki/Technology_evangelist)