Senior HPC & CUDA Performance Engineer - Hybrid/Remote

EPAM Systems; Polska; Hybrid; PLN 280,000 - 420,000; Full time

Benefits offered by this job

Hybrid work model

Remote within Poland

International travel opportunities

Relocation support

Health insurance

Stock purchase plan

Job summary

EPAM Systems is seeking an experienced HPC / CUDA Optimization Engineer to design, develop, and optimize CUDA kernels for large-scale scientific and engineering applications. You will analyze GPU performance with profiling tools, and implement distributed multi-GPU solutions using MPI, NCCL, and NVSHMEM.

The role requires deep CUDA knowledge, strong C++ and Python skills, and the ability to work across cross-functional teams to deliver scalable software.

Qualifications

• 5+ years of C++ and Python development experience.
• Deep knowledge of CUDA, GPU architecture, memory hierarchy, and kernel optimization.
• Experience with GPU profiling and bottleneck analysis.
• Hands-on distributed and multi-GPU computing experience (MPI, NCCL, NVSHMEM).
• Experience developing high-load, performance-critical software.
• Strong debugging and software architecture skills.

Responsibilities

• Design, develop, and optimize CUDA kernels for large-scale applications.
• Analyze and improve GPU performance using profiling tools.
• Implement distributed and multi-GPU solutions with MPI/NCCL/NVSHMEM.
• Develop high-performance C++ and Python code for critical systems.
• Collaborate with cross-functional teams to architect scalable software.
• Debug complex performance and memory issues across GPUs and clusters.
• Apply HPC best practices and numerical methods to improve codebases.

Skills

5+ years C++/Python

CUDA optimization

GPU computing

Performance profiling

MPI/NCCL/NVSHMEM

Distributed systems

Debugging large codebases

Cross-functional collaboration

Tools

MPI

NCCL

NVSHMEM

CUDA toolkit

Job description

EPAM Systems is seeking an experienced HPC / CUDA Optimization Engineer to design, develop, and optimize CUDA kernels for large-scale scientific and engineering applications. You will analyze GPU performance with profiling tools, and implement distributed multi-GPU solutions using MPI, NCCL, and NVSHMEM.

The role requires deep CUDA knowledge, strong C++ and Python skills, and the ability to work across cross-functional teams to deliver scalable software.