The role
- COMP
- $200K - $350K
- EQUITY
- Competitive equity
- LOCATION
- San Francisco
- WORKPLACE
- On-site
- EXPERIENCE
- 2+ years
- VISA
- None, Visa transfers
- STACK
- Python, C/C++
- INDUSTRY
- Robotics, AI, Hardware, Manufacturing
/roles — ROLE_318
Member of Technical Staff, Infrastructure
Early-stage robotics company building two-armed industrial robots and the full stack behind them
The company
Robotics startup building bimanual manipulators for industrial work, owning the hardware, machine learning, and deployment infrastructure needed to make automated physical labor highly reliable.
- STAGE
- Series A-stage
- FUNDING
- $20M+ raised
- TEAM
- ~20 people
- FOUNDED
- 2026
- BACKING
- VC-backed
JD — the work
About the role
You would build and own the infrastructure behind robots already working in the field: large-scale training orchestration, petabyte-level data pipelines, GPU scheduling, and low-latency networking between robots and compute over long distances. The team is small, moves quickly, and runs on a six-day work week. This is not a big-company infrastructure job: you go wherever the current bottleneck is and help define how engineering is done while the code is early enough that those calls will last.
What you'll do
- Own the systems that schedule compute, recover from failures, and coordinate large training jobs across GPU clusters
- Design pipelines that pull petabytes of robot logs and video into training while keeping GPUs saturated
- Close the loop from deployed robots to training, feeding real-world trajectories back in reliably and automatically
- Shift focus week to week, from deployment paths to core libraries to observability gaps
- Create internal tooling whose benefits compound, making engineers and researchers across the team faster
What they're looking for
- 2+ years of experience building distributed systems or ML infrastructure
- Strong Python and C/C++ skills for systems-level work
- Experience with large-scale data pipelines or GPU cluster orchestration
- Comfort owning problems wherever they sit, from deployment tooling to core libraries
- Appetite for six-day weeks on site in San Francisco