Senior LLM Training Performance Engineer (Hybrid)

NVIDIA Corporation; Santa Clara (CA); Hybrid; USD 184,000 - 357,000; Full time

Benefits offered by this job

Equity

Benefits package

Hybrid work environment

Job summary

NVIDIA Corporation seeks a Senior High-Performance LLM Training Engineer to optimize AI training workloads and accelerate performance across thousands of GPUs.

You will profile, analyze, and implement production-quality software across the deep learning stack, from drivers to frameworks, while contributing to MLPerf submissions and architecture studies. Proficiency in PyTorch, CUDA, and C++ is essential.

Qualifications

• PhD in Computer Science, Electrical Engineering or Computer Engineering or MS with extensive experience.
• Strong background in deep learning and neural networks, especially training.
• Proven experience analyzing and tuning application performance; system-level modelling.
• Programming skills in C++, Python, and CUDA.

Responsibilities

• Understand, profile, and optimize AI training workloads on advanced hardware and software platforms.
• Analyze training performance across GPUs and neural networks; prioritize and solve problems.
• Implement production-quality software across NVIDIA's deep learning stack, from drivers to frameworks.
• Build and support MLPerf Training benchmark submissions.
• Develop workloads for future architecture studies in simulators.
• Create tools to automate workload analysis and optimization workflows.

Skills

C++

Python

CUDA

Deep learning

Neural networks

Performance tuning

Education

PhD in Computer Science, Electrical Engineering or Computer Engineering

MS in a related field

Job description

NVIDIA Corporation seeks a Senior High-Performance LLM Training Engineer to optimize AI training workloads and accelerate performance across thousands of GPUs.

You will profile, analyze, and implement production-quality software across the deep learning stack, from drivers to frameworks, while contributing to MLPerf submissions and architecture studies. Proficiency in PyTorch, CUDA, and C++ is essential.