Member of Technical Staff, Training Performance Engineer
Core
Optimizing performance of advanced language models and systems to improve training throughput and accelerator utilization.
Role type
Senior IC machine-learning engineer (training performance)
Builds
High-performant and scalable software for training large language models
Domain
Artificial Intelligence / Natural Language Processing / High-performance computing
Deliverable
production ML models
Required skills
Software engineering, Python, CUDA kernel development, Triton, distributed training strategies, Transformer architectures, ML frameworks (JAX, PyTorch, XLA/MLIR)
Preferred skills
Publications at top-tier venues (NeurIPS, ICML, ICLR, etc.)
Technologies
CUDA, Triton, JAX, PyTorch, XLA, MLIR, Transformers
Responsibilities
Design and write high-performant and scalable software for training; Identify and remove performance bottlenecks; Develop cutting-edge training and profiling tools; Research and experiment with ideas on supercompute and data infrastructure
Seniority
Senior, hands-on IC