Senior Training Infra Engineer - 800+ GPU Scale

Figureai; San Jose (CA); On-site; USD 150,000 - 350,000; Full time

Job summary

Figureai in San Jose, CA is seeking an experienced Training Infrastructure Engineer to manage and optimize their training infrastructure for AI robotics. You will collaborate with AI researchers to scale the training of new model architectures and enhance performance through distributed training algorithms.

A strong background in Python and deep learning frameworks is essential, along with experience managing large-scale GPU clusters. Join an innovative team and help lead the future of humanoid robotics.

Qualifications

• Strong software engineering fundamentals.
• Extensive professional experience with Python and PyTorch.
• Proven track record of scaling and running large-scale training experiments on 800+ GPUs.
• Minimum of 4 years of professional experience in building reliable infrastructure.

Responsibilities

• Design, deploy, and maintain training clusters.
• Optimize and maintain scalable deep learning frameworks.
• Implement distributed training and data loaders.
• Profile, identify, and eliminate training bottlenecks.
• Implement tooling for data processing and model experimentation.

Skills

Python

PyTorch

Backend systems

GPU management

HPC clusters

Education

Bachelor's or Master's degree in Computer Science, Robotics, Engineering, or related field

Tools

AWS

Azure

GCP

CUDA

SLURM

Kubernetes

Job description

Figureai in San Jose, CA is seeking an experienced Training Infrastructure Engineer to manage and optimize their training infrastructure for AI robotics. You will collaborate with AI researchers to scale the training of new model architectures and enhance performance through distributed training algorithms.

A strong background in Python and deep learning frameworks is essential, along with experience managing large-scale GPU clusters. Join an innovative team and help lead the future of humanoid robotics.