Sunnyvale, California, United States 2d ago Not disclosed
Linux HPC & Scientific Computing Jobs
HPC and scientific computing engineers build and operate supercomputing infrastructure and write high-performance simulation code. This track covers cluster administration (Slurm, PBS, Lustre), parallel computing (MPI, OpenMP, CUDA), numerical simulation, and research computing infrastructure. Linux is the universal operating system for HPC. From the Top500 list to university research clusters, every supercomputer runs Linux.
Munich, Bavaria, Germany 2d ago Not disclosed
A
San Jose, California, United States 6d ago $163k – $218k
Long Beach, California, United States May 21 $181k – $249k
Abu Dhabi, United Arab Emirates May 21 Not disclosed
Paris, France May 21 Not disclosed
Singapore, Singapore May 21 Not disclosed
Taipei, Taiwan May 21 Not disclosed
Taipei, Taiwan May 21 Not disclosed
San Jose, California, United States May 21 Not disclosed
San Jose, California, United States May 21 Not disclosed
San Jose, California, United States May 21 Not disclosed
Remote, United States May 21 Not disclosed
San Francisco, California, United States May 21 Not disclosed
Toronto, Ontario, Canada May 21 Not disclosed
San Francisco, California, United States May 21 Not disclosed
San Francisco, California, United States May 21 Not disclosed
PL
Menlo Park, California, United States May 21 $350k – $450k
Pangyo, Gyeonggi, South Korea May 21 Not disclosed
San Francisco, California, United States May 21 $150k – $300k
Frequently Asked Questions
-
High-Performance Computing (HPC) refers to the use of parallel processing across many compute nodes to solve computationally intensive problems: weather simulation, drug discovery, quantum chemistry, computational fluid dynamics, genomics, and astrophysics. HPC systems range from university clusters to national laboratory supercomputers.
-
HPC Linux skills include cluster administration (Slurm, PBS Pro, LSF), high-speed networking (InfiniBand, HDR, RoCE), parallel file systems (Lustre, GPFS), module systems (Lmod), MPI installation and tuning, NUMA topology awareness, and kernel tuning for latency-sensitive workloads. Scripting (Bash, Python) for job submission and workflow automation is essential.
-
Fortran remains dominant in legacy scientific codes (climate models, legacy CFD). C and C++ are used for performance-critical applications. Python is the scripting and analysis language of choice. For GPU computing, CUDA (NVIDIA) and HIP (AMD) are used. Modern scientific frameworks increasingly use Python with C++ extensions (NumPy, SciPy, PyTorch).
-
National laboratories (DOE labs: Argonne, Oak Ridge, LLNL, Sandia), government research agencies (NSF, NIH), university research computing centres, pharmaceutical companies (Pfizer, Novartis), aerospace and defence (Boeing, Lockheed), financial services for quantitative modelling, and cloud providers building HPC offerings.