Research Scientist: Pretraining

ML Engineer — Robot Foundation Model Pretraining

We are building large-scale embodied intelligence systems designed to operate in complex real-world environments. Our work focuses on training robot foundation models using massive multimodal datasets spanning video, language, proprioception, action traces, and related sensor modalities.

We are seeking ML Engineers to design and execute large-scale pretraining efforts that develop general-purpose capabilities across embodiments, tasks, and environments. This role focuses on turning raw robotic interaction data into scalable, generalizable intelligence.

What You’ll Do

Design and Execute Large-Scale Pretraining

• Lead or contribute to large-scale training runs for robot foundation models
• Work with transformer-based and diffusion-based architectures for multimodal learning
• Define model architectures, training objectives, and curricula for multimodal robotic data

Develop Data and Training Strategies

• Design scalable data mixtures and sampling strategies across petabyte-scale datasets
• Structure training curricula across vision, action, state, language, and other modalities
• Convert raw robotic interaction data into usable signals for model training

Improve Model Performance Through Analysis

• Run ablation studies to understand scaling laws, data quality effects, and architecture tradeoffs
• Analyze training dynamics and large-model failure modes
• Iterate on data and training design to improve generalization and robustness

Collaborate on Large-Scale Systems

• Work closely with ML infrastructure and systems teams to improve cluster utilization, throughput, and reliability
• Contribute to multi-node, multi-GPU distributed training efforts
• Ensure efficient execution of large-scale training workflows

Shape Data Collection and Model Direction

• Guide data collection efforts toward high-impact capabilities and gaps
• Identify and integrate new datasets to improve model coverage and generalization
• Bridge raw data generation with downstream model performance

What We’re Looking For

• Experience training large-scale transformer or diffusion models (e.g., language, video, audio, or generative models)
• Experience with multi-node, multi-GPU distributed training systems
• Strong understanding of:
• Scaling laws
• Optimization dynamics
• Large-model training behavior and failure modes
• Strong proficiency in PyTorch and ability to debug across the full ML stack
• Comfort working with large-scale datasets and rapid experimentation
• Strong empirical rigor combined with fast iteration speed

Preferred Experience

• Experience with multimodal or generative model training at scale
• Background in robotics, embodied AI, or sequential decision-making systems
• Experience with large distributed training infrastructure
• Familiarity with curriculum learning or dataset mixture design
• Experience analyzing large-scale training runs and extracting actionable insights

Why This Role Matters

• Build the core intelligence layer for general-purpose robotic systems
• Directly shape how robots learn from large-scale real-world data
• Work at the intersection of large-scale AI training, systems, and robotics
• Develop foundation models that generalize across embodiments and environments

About the Company

We are a research-driven AI and robotics company focused on building scalable embodied intelligence systems. By combining advances in machine learning, large-scale training infrastructure, and robotics, we aim to develop systems capable of operating robustly in the physical world.

We are committed to building an inclusive and diverse workplace and encourage applicants from all backgrounds to apply.