Research Scientist: Scalable General Models for Robots

pantograph.com; San Francisco (CA); On-site; USD 150,000 - 230,000; Full time

Job summary

Pantograph in San Francisco is seeking research scientists to scale simple methods across large video and multimodal datasets, with applications to robotics. You will work on training regimes that generalize from internet-scale video to real-world robots and fleets.

We value candidates who have built models on large GPU clusters, are comfortable with Kubernetes, and favor straightforward, scalable approaches with clear, measurable metrics.

Qualifications

• Experience with large-scale pre-training across video, multimodal, image, or language.
• Experience in self-supervised, goal-conditioned, or unsupervised RL.
• Background in robotics models trained on large-scale data.
• Experience with large GPU clusters and Kubernetes.
• Preference for simple, scalable methods with clear metrics.

Responsibilities

• Scale simple methods across large datasets and domains.
• Collaborate on video, real-world video, and robotics projects.
• Develop and evaluate expressive metrics to measure improvements.
• Share results through open-source contributions or publications.

Skills

Large-scale pre-training

Self-supervised RL

Robotics models

Video sequence modeling

Scalable methods

Expressive metrics

Tools

Kubernetes

JAX

Job description

Pantograph in San Francisco is seeking research scientists to scale simple methods across large video and multimodal datasets, with applications to robotics. You will work on training regimes that generalize from internet-scale video to real-world robots and fleets.

We value candidates who have built models on large GPU clusters, are comfortable with Kubernetes, and favor straightforward, scalable approaches with clear, measurable metrics.