Teams at Pantograph

Implement and scale research ideas into large-scale training systems: run experiments on GPU clusters, build data and training infrastructure, process massive multimodal datasets, improve observability, and bridge research code with production systems for robotics and multimodal model development.

Build and ship robot software features, develop internal test infrastructure and tooling, work closely with hardware teams and customers, and improve reliability and user experience for fleet robots.

Research scientist to scale simple, general models trained on internet-scale video and robot data. Develop and pre-train large multimodal/video models, apply self-supervised or unsupervised RL, train on large GPU clusters, and measure model performance rigorously.