High-Performance LLM Training Engineer
NVIDIA
About the Role
NVIDIA is seeking engineers specializing in performance analysis and optimization to improve the efficiency of LLM training workloads. This position focuses on optimizing NVIDIA's high-performance LLM software stack in frameworks like PyTorch and JAX for high-performance pre-training and post-training on thousands of GPUs, while also helping shape hardware roadmaps for the next generation of GPUs powering the AI revolution.
Key Responsibilities
- Understand, analyze, profile, and optimize AI training workloads on innovative hardware and software platforms.
- Understand the big picture of training performance on GPUs, prioritizing and then solving problems across all state-of-the-art LLMs.
- Implement production-quality software in multiple layers of NVIDIA's deep learning platform stack.
- Implement key LLM training workloads in NVIDIA's proprietary processor and system simulators to enable future architecture studies.
- Build agentic tools to automate workload analysis, workload optimization, and other critical workflows.
Requirements & Tech Stack
- Programming skills in C++, Python, and CUDA.
- Familiarity with the fundamentals of GPU architecture.
- Experience with deep learning frameworks like PyTorch and JAX.
Qualifications
- M.S. or PhD in Computer Science, Electrical Engineering or Computer Engineering (or equivalent experience).
- Strong background in deep learning and neural networks, in particular model training optimizations.
- A deep background in computer architecture.
- Proven experience analyzing and tuning application performance & processor and system-level performance modeling.