Sr. HPC Engineer - Platforms
PsiQuantum16 hours ago
Responsibilities
- Design, implement, and optimize HPC software using NVIDIA GPUs and AMD CPUs.
- Apply MPI, CUDA, OpenMP, and other parallel-computing techniques across multi-core and multi-GPU systems.
- Design, deploy, scale, and optimize HPC nodes and cluster configurations for growing computational workloads.
- Identify and implement algorithms suited to GPU and CPU computing while maintaining performance as node counts increase.
- Profile and benchmark HPC applications using NVIDIA Nsight, AMD μProf, and Intel VTune to identify bottlenecks.
- Collaborate with scientists, engineers, and IT staff to optimize applications for research and production environments.
- Document code and scaling processes and report performance metrics, improvements, and results.
Requirements
- Bachelor's or master's degree in computer science, electrical engineering, computational science, or a related field.
- 3+ years of HPC development experience focused on parallel and distributed computing and scalable systems.
- 5+ years of scientific software development experience.
- Proficiency in CUDA and familiarity with NVIDIA's HPC software stack, including cuDNN, NCCL, and TensorRT.
- Expertise in MPI, OpenMP, and other parallel programming paradigms.
- Strong understanding of cluster scaling and distributed, multi-node software optimization.
- Strong proficiency in C/C++, Python, and Ansible, Terraform, or Packer.
- Familiarity with NVIDIA Nsight, AMD μProf, and Intel VTune.
- Ability to work independently and collaboratively with strong problem-solving, communication, and attention-to-detail skills.
- Preferred: experience with Slurm, PBS, or other HPC scheduling and resource-management tools.
- Preferred: knowledge of Singularity or Docker in an HPC context.
- Preferred: understanding of cloud-based HPC scaling solutions or hybrid HPC environments.
- Preferred: experience with GPU-accelerated machine-learning frameworks such as TensorFlow, PyTorch, or PySCF.
About PsiQuantum
PsiQuantum builds photonic quantum computers designed to achieve fault-tolerant, million‑qubit scale, and develops algorithms and software for commercial use. The privately held company is headquartered in Palo Alto, and in 2024 announced government-funded projects to build utility-scale systems in Brisbane and Chicago. It collaborates with enterprise and defense customers, with publicly named partners including Lockheed Martin, Mercedes‑Benz, Boehringer Ingelheim, and Mitsubishi Chemical.