Software Engineer: ML Infra
About The Position
Generalist builds embodied foundation models for the physical world, with an initial focus on dexterity. The company develops data, models, and hardware to enable robots to interact intelligently with the physical world. The Software Engineer: ML Infra will support the training of very large robot foundation models using large fleets of current-generation GPU hardware and infrastructure. The role includes building and optimizing infrastructure for distributed training and researcher experiments, with particular attention to storage and data loading at scale. The position also owns inference infrastructure for robot fleets: on-premises GPUs attached to robots operating with demanding real-time and latency budgets in compute-constrained environments.
Requirements
- Experience managing large fleets of GPUs for large-scale, long-term, highly distributed training runs or inference.
- Deep experience with Slurm or Kubernetes for ML workload orchestration.
- Experience building high-scale ML data loaders and preparation systems.
- Deep understanding of the layers of ML hardware, storage, and networking stacks.
- Experience in the Nvidia GPU ecosystem.
Responsibilities
- Own GPU compute fleets.
- Ensure GPUs are easy for researchers to use and are maximally utilized.
- Optimize and improve ML data loading, transport, and storage in highly distributed, fully utilized environments.
- Orchestrate robot inference fleets.
Stand Out From the Crowd
Upload your resume and get instant feedback on how well it matches this job.
What This Job Offers
Job Type
Full-time
Education Level
No Education Listed
