NVIDIA is seeking an experienced HPC engineer to own and optimize the federated LSF scheduler across 15-25 cells, tuning mbatchd/mbs chd, and analyzing scheduling cycles as the farm scales. You will address contention patterns and design cell topology for growth.
Required are a BS/MS in CS/CE, 8+ years in HPC with 5+ years on IBM Spectrum LSF, deep LSF internals debugging, and robust Linux scripting (Python/Perl/shell). Hybrid work with equity is included.