Robotics AI Research Scientist — Large-Scale Foundation Models

Generalist AI; Boston (MA); On-site; USD 180,000 - 280,000; Full time

Job summary

Generalist is building the base intelligence layer for robotics by pretraining large-scale robot foundation models from multimodal data. You will design and run core training efforts to enable new general capabilities across embodiments and environments.

If you have worked on pretraining large-scale ML systems or generative foundation models (multimodal, language, audio), this role may be a fit. You’ll collaborate with ML Infra and Systems to push performance.

Qualifications

• Experience with pretraining large-scale ML systems or generative foundation models (multimodal, language, audio) is a plus.
• Strong background in transformer or diffusion models used at scale for generative tasks.
• Experience leading multi-node, multi-GPU distributed training efforts.
• Knowledge of scaling laws and data quality effects on model performance.
• Solid PyTorch fundamentals and debugging across stack layers.

Responsibilities

• Designing and executing large-scale pretraining runs for robot foundation models (transformer- and diffusion-based).
• Defining model architectures, objectives, and training curricula across multimodal robotic data.
• Developing scalable data mixtures and sampling strategies over petabyte-scale datasets.
• Guiding data collection directions and sourcing new datasets for training.
• Running ablations to understand scaling laws and architecture tradeoffs.
• Collaborating with ML Infra and Systems to improve cluster utilization and reliability.
• Turning raw robotic interaction data into generalizable model capabilities.

Skills

Large-scale ML systems

Distributed training

PyTorch

Multimodal models

Job description

Generalist is building the base intelligence layer for robotics by pretraining large-scale robot foundation models from multimodal data. You will design and run core training efforts to enable new general capabilities across embodiments and environments.

If you have worked on pretraining large-scale ML systems or generative foundation models (multimodal, language, audio), this role may be a fit. You’ll collaborate with ML Infra and Systems to push performance.