Senior High-Performance LLM Training Engineer

NVIDIA•Santa Clara, CA

10d•$184,000 - $356,500•Hybrid

About The Position

NVIDIA is seeking an experienced engineer specializing in performance analysis and optimization to improve the efficiency of LLM training workloads. The role optimizes NVIDIA’s high-performance LLM software stack in frameworks such as PyTorch and JAX for training across thousands of GPUs, and helps shape hardware roadmaps for future GPUs. The engineer works across NVIDIA’s hardware and software stack, from GPU architecture to application code, in an environment that encourages innovation and collaboration.

Requirements

PhD in Computer Science, Electrical Engineering, or Computer Engineering and 5+ years of experience, or a master’s degree (or equivalent experience) and 8+ years of meaningful work experience.
Strong background in deep learning and neural networks, particularly training.
Deep background in computer architecture and familiarity with GPU architecture fundamentals.
Proven experience analyzing and tuning application performance, as well as processor- and system-level performance modeling.
Programming skills in C++, Python, and CUDA.

Responsibilities

Understand, analyze, profile, and optimize AI training workloads on innovative hardware and software platforms.
Analyze training performance on GPUs, prioritize performance issues, and solve problems across state-of-the-art neural networks.
Implement production-quality software across multiple layers of NVIDIA’s deep learning platform stack, from drivers to deep learning frameworks.
Build and support NVIDIA submissions to the MLPerf Training benchmark suite.
Implement key deep learning training workloads in NVIDIA’s proprietary processor and system simulators to support future architecture studies.
Build tools to automate workload analysis, workload optimization, and other critical workflows.