Senior GPU Software Engineer - CUTLASS & MLIR

NVIDIA; Hillsboro (OR); On-site; USD 184,000 - 287,000; Full time

Benefits offered by this job

Equity

Benefits

Job summary

NVIDIA is seeking an expert to design abstractions for Tensor Core and GPU hardware features within MLIR, Python, and C++. You will develop CUTLASS components, contribute to the MLIR-based compiler stack, and write example kernels to showcase high-performance GPU capabilities.

You will collaborate with GPU architecture, CUDA, and NVVM/ PTX teams to shape future hardware features while applying your deep knowledge of parallelism and computer architecture.

Qualifications

• Masters or PhD in Computer Science, Computer Engineering, or related field (or equivalent).
• 3+ years of relevant industry experience.
• Strong proficiency in C++ programming and software design, including debugging and testing.
• Experience with high-performance code generation and compiler transformations/optimizations.
• Deep understanding of computer architecture and parallel computing models.

Responsibilities

• Develop core components of the CUTLASS platform using CUDA C++ and CUTLASS Python DSL.
• Advance the MLIR-based backend compiler stack by designing dialects and passes.
• Author example kernels to demonstrate GPU features for high performance.
• Collaborate with GPU architecture, CUDA, and NVVM/ PTX teams to provide feedback and assess future features.

Skills

C++

CUDA

Python

MLIR

GEMM

Performance tuning

Education

Master's or PhD

Tools

CUTLASS

NVVM/PTX

Job description

NVIDIA is seeking an expert to design abstractions for Tensor Core and GPU hardware features within MLIR, Python, and C++. You will develop CUTLASS components, contribute to the MLIR-based compiler stack, and write example kernels to showcase high-performance GPU capabilities.

You will collaborate with GPU architecture, CUDA, and NVVM/ PTX teams to shape future hardware features while applying your deep knowledge of parallelism and computer architecture.