Robotics AI Foundation Scientist

Generalist; Somerville (NJ); On-site; USD 180,000 - 240,000; Full time

Job summary

Generalist in New Jersey is building the base intelligence layer for robotics, training large-scale robot foundation models from multimodal data. You will design and run core training efforts that enable models to acquire general capabilities across embodiments, tasks, and environments.

If you have led large-scale ML systems and multi-node training, this role will suit you; you’ll collaborate with ML Infra and Systems, turning raw robotic interaction data into robust, scalable intelligence.

Qualifications

• Experience with large-scale ML systems at senior level (L6+).
• Extensive experience training transformer and diffusion models at scale.
• Led multi-node, multi-GPU distributed training efforts.
• Strong PyTorch fundamentals and debugging across stack.
• Interest in building general-purpose robot intelligence from first principles.

Responsibilities

• Design and execute large-scale pretraining runs for robot foundation models (transformer- and diffusion-based).
• Define model architectures, objectives, and training curricula across multimodal data.
• Develop scalable data mixtures and sampling strategies for petabyte-scale datasets.
• Guide data collection directions and sourcing of new datasets.
• Run ablations to understand scaling laws and data quality effects.
• Collaborate with ML Infra and Systems to improve cluster utilization and reliability.
• Turn raw robotic interaction data into generalizable model capabilities.

Skills

Large-scale ML

Transformer models

Diffusion models

Distributed training

PyTorch

Data sampling

Empirical rigor

Job description

Generalist in New Jersey is building the base intelligence layer for robotics, training large-scale robot foundation models from multimodal data. You will design and run core training efforts that enable models to acquire general capabilities across embodiments, tasks, and environments.

If you have led large-scale ML systems and multi-node training, this role will suit you; you’ll collaborate with ML Infra and Systems, turning raw robotic interaction data into robust, scalable intelligence.