Member of Technical Staff: Infrastructure
Pantheon · San Francisco
Roboticsonsite$200k–$350k
At a glance
Highlights
- Own training infrastructure powering large-scale video and multimodal models.
- Build robots for manufacturing toward a software-defined physical world.
- Debug across model, infrastructure and hardware to keep research moving.
What they offer
- Equity
Ideal candidate
Infrastructure-minded engineer who loves making research teams fast and trusts reliable systems. You thrive debugging across hardware and training code to remove bottlenecks at scale.
Requirements
- Extensive experience with PyTorch or JAX in production or research environments
- Built or operated distributed training systems across multi-GPU or multi-node setups
- Deep understanding of training performance bottlenecks around data loading, memory, compute, and GPU utilization
- Experience building or optimizing large-scale data pipelines, ideally for video, multimodal, or foundation model training data
- Skills to debug across training code, infrastructure, and hardware
- Deep engineering judgment and bias toward reliable, fast, reusable systems trusted by research teams
Lab profile
About Pantheon
Building billions of useful robots, starting with bi-manual manipulators for manufacturing.
What they build
- Bi-manual manipulators for manufacturing
Known for
- Bimanual manipulation
- Full-stack deployable robotics
Posted · Last seen