High Performance Computing Engineer
Polyhedra Network
Other Engineering
San Francisco, CA, USA
Posted on Mar 25, 2025
About Polyhedra Network
We are pioneering advanced computational solutions requiring massive computational power and sophisticated optimization. Our HPC team designs and implements cutting-edge solutions that push the boundaries of computational performance.
About The Role
We're seeking a High Performance Computing Engineer with deep expertise in parallel computing and heterogeneous architecture optimization. The ideal candidate will have extensive experience with CUDA/ROCm, MPI, and large-scale system optimization.
AI Tool Proficiency Requirements
- Advanced expertise in using AI coding assistants for HPC code generation and optimization
- Demonstrated ability to leverage AI tools for performance debugging and analysis
- Experience using AI systems for parallel algorithm implementation and optimization
- Strong proficiency in using AI to accelerate development workflows
- Ability to effectively combine AI-generated solutions with manual optimization
- Experience using AI tools for code review and quality assurance
What You Will Do
- Design and optimize high-performance computing applications for large-scale clusters
- Implement and tune parallel algorithms across distributed systems
- Develop efficient CUDA/ROCm implementations for computational kernels
- Create and maintain MPI-based distributed computing solutions
- Optimize memory access patterns and communication protocols
- Profile and analyze performance bottlenecks in distributed systems
- Implement efficient linear algebra and scientific computing routines
- Manage and optimize job scheduling and resource allocation
- Design and implement fault tolerance and recovery systems
- Collaborate with research teams to optimize computational workloads
Required Qualifications
- M.S. or Ph.D. in Computer Science, Engineering, or related field from a top-tier university
- Expert knowledge of parallel programming models (MPI, OpenMP)
- Advanced expertise in CUDA or ROCm programming
- Strong proficiency in C++ and parallel algorithm implementation
- Deep understanding of computer architecture and memory hierarchies
- Expert-level knowledge of Linux/Unix environments
- Experience with large-scale cluster management
Preferred Qualifications
- Experience with InfiniBand and high-speed interconnects
- Knowledge of scientific computing libraries (BLAS, LAPACK)
- Expertise in vectorization and SIMD optimization
- Background in distributed algorithms
- Experience with performance modeling and prediction
- Familiarity with job scheduling systems (Slurm, PBS)
- Experience with container technologies for HPC
- ACM-ICPC Regional or World Finals medalist
- USACO (USA Computing Olympiad) Gold/Platinum award
- Top-tier algorithmic competition achievements
Technical Skills
- Advanced parallel programming techniques
- Proficiency in performance optimization tools
- Strong debugging and profiling skills
- Expertise in distributed computing concepts
- Knowledge of network topology and optimization
- Experience with scientific computing applications
Required Tools & Technologies
- CUDA/ROCm
- MPI
- OpenMP
- Linux/Unix
- Performance profiling tools
- Job scheduling systems
- Version control systems
- C/C++
What We Offer
- Access to cutting-edge HPC infrastructure
- Competitive compensation package
- Professional development opportunities
- Collaboration with leading researchers
- Health and retirement benefits