Lead HPC / CUDA Optimization Engineer

EPAM Systems; Stati Uniti; Remoto; USD 180.000 - 240.000; Tempo pieno

Descrizione del lavoro

EPAM Systems Inc. seeks an experienced Lead HPC / CUDA Optimization Engineer to drive high-performance GPU software for large-scale scientific and engineering applications. You will code hands-on and mentor engineers while shaping architecture and performance decisions.

You will optimize CUDA kernels, profile GPU performance, and collaborate with cross-functional teams to deliver scalable, memory-efficient solutions across distributed systems. Excellent English required.

Competenze

• 8+ years of professional experience with C++ and Python development.
• Deep knowledge of CUDA, GPU architecture, PTX, memory hierarchy, and kernel optimization.
• Experience with GPU profiling, performance analysis, and bottleneck identification.
• Experience developing and optimizing performance-critical, high-load software systems.
• Strong debugging and software architecture skills.
• Proven ability to lead technical initiatives and collaborate with cross-functional teams.
• English proficiency at B2 level or higher.

Mansioni

• Design, develop, and optimize CUDA kernels for large-scale scientific and engineering applications.
• Analyze and improve GPU performance using profiling tools to identify and resolve bottlenecks.
• Develop and maintain high-performance C++ and Python code for performance-critical, high-load systems.
• Collaborate with cross-functional teams to architect scalable and efficient software solutions.
• Debug complex performance and memory issues across GPU architectures and distributed systems.
• Research and apply best practices in HPC, numerical methods, and GPU computing to improve existing codebases.
• Provide technical leadership and guidance on GPU optimization and software architecture.
• Drive engineering best practices and contribute to technical decision-making for HPC and GPU-accelerated solutions.

Conoscenze

C++ development

Python development

CUDA programming

GPU optimization

Team leadership

Strumenti

MPI

NCCL

NVSHMEM

HDF5

NVIDIA Nsight

Descrizione del lavoro

We are seeking an experienced Lead HPC / CUDA Optimization Engineer with strong expertise in modern C++, CUDA, GPU computing, and performance optimization for large-scale scientific and engineering applications. In this role, you will remain highly hands-on while also providing technical leadership, mentoring engineers, and helping drive architectural and performance-related decisions across the team.

Responsibilities

• Design, develop, and optimize CUDA kernels for large-scale scientific and engineering applications
• Analyze and improve GPU performance using profiling tools to identify and resolve bottlenecks
• Develop and maintain high-performance C++ and Python code for performance-critical, high-load systems
• Collaborate with cross-functional teams to architect scalable and efficient software solutions
• Debug complex performance and memory issues across GPU architectures and distributed systems
• Research and apply best practices in HPC, numerical methods, and GPU computing to improve existing codebases
• Provide technical leadership and guidance on GPU optimization, performance engineering, and software architecture
• Drive engineering best practices and contribute to technical decision-making for HPC and GPU-accelerated solutions

Requirements

• 8+ years of professional experience with C++ and Python development
• Deep knowledge of CUDA, GPU architecture, PTX, memory hierarchy, and kernel optimization
• Experience with GPU profiling, performance analysis, and bottleneck identification
• Experience developing and optimizing performance-critical, high-load software systems
• Strong debugging and software architecture skills
• Proven ability to lead technical initiatives and collaborate effectively with cross-functional teams
• English proficiency at B2 level or higher
• Nice to have Hands-on experience with distributed and multi-GPU computing (MPI, NCCL, NVSHMEM)
• Experience with RTM/FWI, seismic imaging, or geoscience applications
• Development of HDF5-based data viewing and analysis tools
• Familiarity with scientific computing and numerical methods