Remote Senior Training Infrastructure Engineer—Multi-GPU AI

Luma AI; San Francisco (CA); Hybrid; USD 187,500 - 395,000; Full time

Job summary

Luma AI in SF Bay Area is hiring for a Research Scientist/Engineer in Training Infrastructure. You will design distributed training systems for thousands of GPUs and enable researchers to push multimodal foundation models.

You should have deep experience with PyTorch, CUDA and distributed systems, plus Linux and scripting. The role offers remote/international options including London; hybrid/remote work supported.

Qualifications

• Experience with distributed PyTorch training on GPUs.

Responsibilities

• Design, implement, and optimize distributed training systems for models with thousands of GPUs
• Research and implement advanced parallelization techniques (FSDP, Tensor Parallel, Pipeline Parallel, Expert Parallel)
• Build monitoring, visualization, and debugging tools for large-scale training runs
• Optimize training stability, convergence, and resource utilization across massive clusters

Skills

Large-scale GPU training

Distributed PyTorch training

GPU clusters

Linux scripting

Containerization

Cloud infrastructure

NCCL

MPI

CUDA

Tools

PyTorch

CUDA

Distributed systems

NCCL

MPI

Linux

Scripting

Containerization

Cloud infra

Job description

Luma AI in SF Bay Area is hiring for a Research Scientist/Engineer in Training Infrastructure. You will design distributed training systems for thousands of GPUs and enable researchers to push multimodal foundation models.

You should have deep experience with PyTorch, CUDA and distributed systems, plus Linux and scripting. The role offers remote/international options including London; hybrid/remote work supported.