SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Luma AI is seeking a Research Scientist / Engineer to build and optimize the distributed systems that train Luma's large-scale multimodal models across thousands of GPUs. This role focuses on enabling researchers to innovate on top of reliable, efficient, and scalable infrastructure.
You will own the design, implementation, and optimization of distributed training systems for foundation models at massive scale. This is deep PyTorch, CUDA, and distributed-systems work requiring expertise in advanced parallelism techniques, training stability, and cluster utilization. The role is suited for an engineer with proven experience solving real problems in foundation-model training at scale.
Key Responsibilities:
- Design, implement, and optimize efficient distributed training systems for models across thousands of GPUs
- Research and implement advanced parallelization strategies (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
First 90 Days:
Days 1–30: Immerse in the current training stack and diagnose where stability and utilization challenges exist at scale.
Days 30–60: Ship and validate a parallelization or stability improvement that measurably enhances a real training run.
Days 60–90: Build monitoring and tooling infrastructure to keep large runs reliable and efficient.
About Luma: Luma's mission is to build unified general intelligence that can generate, understand, and operate in the physical world. The company believes multimodality is critical for intelligence, with vision being the next frontier beyond language models.
Requirements:
- Extensive experience with distributed PyTorch training and parallelisms in foundation-model training
- Deep understanding of GPU clusters, networking, and storage systems
- Familiarity with communication libraries (NCCL, MPI) and distributed-system optimization
- Must have worked at the level of FSDP and multi-node training
Nice to Have:
- Strong Linux systems administration and scripting
- Experience managing training runs across 100+ GPUs
- Experience with containerization, orchestration, and cloud infrastructure
