Infrastructure & Systems Lead

Pantograph; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Pantograph is building a scalable infrastructure stack for a fleet of robots, streaming large video data to training clusters, with frequent model updates and strict real-time latency goals.

The role requires owning the end-to-end system, including networking, storage, databases, embedded software, and deployment pipelines, and optimizing GPU performance across the stack.

Qualifications

• Designed and operated large-scale distributed systems from scratch.
• Managed fleets of hundreds or thousands of computers.
• Deep experience with GPU performance optimization, including CUDA kernels.
• Worked on real-time embedded systems or robotics infrastructure.
• Built petabyte-scale storage and database systems.
• Experience with high-performance networking and video encoding pipelines.

Responsibilities

• Architect and own the infrastructure pipeline behind a fleet of robots and inference clusters.
• Touch networking, storage, databases, embedded software, deployment systems, and GPU optimization.
• Own the architecture decisions that tie these components together to meet real-time latency budgets.

Skills

Distributed systems design

Fleet management

GPU performance optimization

Real-time embedded systems

Petabyte-scale storage

High-performance networking

Video encoding pipelines

Tools

CUDA kernels

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 games, real-world video, and our own fleet of affordable, durable robots.

We're looking for someone to architect and own the entire infrastructure pipeline behind that fleet: thousands of robots with embedded GPUs, communicating over wifi to inference clusters, streaming tens of petabytes of video to training clusters, with new model weights deployed every few minutes — all operating within strict real-time latency budgets.

This role requires someone who can hold an entire system in their head and optimize it end-to-end. You'll touch networking, storage, databases, embedded software, deployment systems, and GPU optimization — and you'll own the architecture decisions that tie them together.

You might be a good fit if you have:

• Designed and operated large-scale distributed systems from scratch

• Managed fleets of hundreds or thousands of computers

• Deep experience with GPU performance optimization, including writing custom CUDA kernels

• Worked on real-time embedded systems or robotics infrastructure

• Built petabyte-scale storage and database systems

• Experience with high-performance networking and video encoding pipelines

Nice to have:

• Rust and low-level performance optimization experience

• Experience taking infrastructure from prototype to production in a small team

We care much more about what you've built than any specific credential. We're a small, fast-moving team working together in person in San Francisco. If you're excited about architecting novel systems at unprecedented scale, we'd love to talk.