Computer Science
[Open soon]
Work description
This grant aims to extend llama.cpp to support efficient distributed inference on multi-node ARM systems, leveraging RDMA (Remote Direct Memory Access). The goal is to reduce communication overhead and improve scalability across ARM nodes, which are increasingly relevant in energy-efficient HPC environments and edge computing.
Academic Qualifications
Bachelor’s degree or enrollment in a Master’s program in Computer Engineering, High-Performance Computing, or Computer Science.
Minimum profile required
- Practical experience with Linux systems (basic administration, Bash/Python scripting);- Knowledge of parallel programming (MPI, OpenMP, or CUDA/HIP);- Experience with SLURM or another HPC workload scheduler;- Knowledge of AI/ML frameworks (PyTorch or TensorFlow);- Fluency in English (written and spoken), essential for communication within the European consortium.
Preference factors
- Experience with HPC software management tools (EasyBuild, Spack, EESSI); - Experience with containerization in HPC environments (Singularity/Apptainer); - Experience developing RDMA-based communication stacks; - Previous participation in hackathons, bootcamps, or technical training events.
Application Period
Since 14 May 2026 to 27 May 2026
[Open soon]
Centre
High-Assurance Software