CUTLASS GPU Performance Engineer – Kernel Benchmarking

NVIDIA; Hillsboro (OR); On-site; USD 152,000 - 287,500; Full time

Job summary

NVIDIA is seeking a skilled engineer to join the CUTLASS team in Hillsboro, Oregon, focusing on performance benchmarking of deep learning models. You'll work on optimizing performance, identifying gaps, and automating processes to enhance CUTLASS kernel performance.

The ideal candidate has a Master's or PhD, strong programming skills in Python and C++, and over 3 years of relevant experience. NVIDIA offers a competitive salary ranging from 152,000 to 287,500 USD based on experience and level, with additional equity and benefits.

Qualifications

• 3+ years of relevant industry experience.
• Experience with performance benchmarking of deep learning frameworks.
• Hands-on experience with developing performance models.

Responsibilities

• Benchmark the performance of deep learning models’ inference and training.
• Identify gaps between theoretical and realized performance and suggest improvements.
• Develop tooling for automating performance optimization.

Skills

Strong programming skills in Python

Strong programming skills in C++

Software performance analysis and optimization

Understanding of computer architecture

Education

Masters or PhD degree in Computer Science, Computer Engineering, or related field

Tools

PyTorch

JAX

TRT-LLM

Job description

NVIDIA is seeking a skilled engineer to join the CUTLASS team in Hillsboro, Oregon, focusing on performance benchmarking of deep learning models. You'll work on optimizing performance, identifying gaps, and automating processes to enhance CUTLASS kernel performance.

The ideal candidate has a Master's or PhD, strong programming skills in Python and C++, and over 3 years of relevant experience. NVIDIA offers a competitive salary ranging from 152,000 to 287,500 USD based on experience and level, with additional equity and benefits.