Principal High-Performance LLM Training Engineer

What you'll do

• Optimize LLM workloads across compute, memory, communication, frameworks, and system scaling; develop tools, benchmarks, and performance models.
• Translate workload findings into hardware and software recommendations and influence architecture and roadmap decisions across teams.
• Provide technical leadership and mentor engineers in large-scale AI performance analysis and optimization.

What they ask for

• MS or PhD (or equivalent experience) in Computer Science, Electrical Engineering, Computer Engineering, or a related field.
• 12+ years of relevant work or research experience.
• Principal-level technical impact in a relevant area, with hands-on experience optimizing large-scale deep learning workloads, especially transformer-based models and LLM training or post-training.
• Strong GPU architecture knowledge and experience with distributed training, performance profiling, benchmarking, and modeling.

Pay

272,000 USD - 431,250 USD (base salary)

Tools

NVIDIA GPUs, PyTorch, JAX, NeMo, NeMo RL, CUDA

AI focus

Optimize LLM pre-training and post-training, including reinforcement learning and fine-tuning, and guide AI hardware and software architecture using workload insights.