Research Engineer: Scale-Driven AI Systems & Robotics

Pantograph; San Francisco (CA); On-site; USD 180,000 - 230,000; Full time

Job summary

Pantograph is training general models that start by watching internet-scale video and end up on robots. We think the path to capable robots runs through general intelligence rather than narrow, robot-specific skills.

We're scaling simple methods across video, real-world data, and our fleet of durable robots. You’ll work across the boundary between research and engineering: implementing new ideas, scaling experiments across large GPU clusters, building the systems that let us iterate quickly, and

Qualifications

• Experience training models on large GPU clusters.
• Proficiency with Kubernetes and distributed systems.
• Ability to scale experiments across multiple clusters and datasets.
• Track record building production-like ML systems with observable metrics.

Responsibilities

• Design scalable training pipelines and multimodal models.
• Scale experiments across large GPU clusters and data stores.
• Build systems enabling rapid iteration between research and production.
• Analyze failures, improve observability, and drive system reliability.
• Collaborate across research and engineering to deploy capabilities.

Skills

Kubernetes

Distributed systems

Large-scale data processing

Multimodal model training

Experimentation & debugging

Research-to-product bridge

Python / ML engineering

Tools

JAX

CUDA kernels

Linux kernel programming

GPU clusters

TensorFlow / PyTorch

Job description

Pantograph is training general models that start by watching internet-scale video and end up on robots. We think the path to capable robots runs through general intelligence rather than narrow, robot-specific skills.

We're scaling simple methods across video, real-world data, and our fleet of durable robots. You’ll work across the boundary between research and engineering: implementing new ideas, scaling experiments across large GPU clusters, building the systems that let us iterate quickly, and