Machine Learning Infrastructure Engineer

About the company

Our client is a robotics company.

The role

Raydar is recruiting for this role on behalf of our client. Develop and maintain the backbone of large-scale model training, handling job dispatch, resource allocation, data flow and communication between machines. The role suits an engineer who will move to whichever area currently limits the team and help set engineering standards in a small, fast-moving group.

What you'll do

- Build and maintain the layers that dispatch training jobs, allocate hardware, recover from failures and connect machines.

- Design large-scale data pipelines that keep accelerators supplied with high volumes of operational data and video.

- Develop automated feedback loops that move production data back into model training.

- Shift focus to whichever area is the current bottleneck, from deployment paths to shared libraries and observability.

- Build internal tooling that speeds up engineers and researchers.

Requirements

What we're looking for

- At least 2 years of engineering experience.

- Currently works at a leading technology company or research lab with a high hiring bar, ranked above large technology firms.

- Practical experience running machine learning workloads on distributed systems, including multi-node GPU training, PyTorch, or GPU scheduling and networking. Data infrastructure or backend work without ML training exposure does not qualify.

- Deep foundations in building reliable software at scale, spanning distributed computing, data flow or workload management. Kernel-level or inference tuning alone does not meet the bar.

- Able to work on-site in San Francisco six days a week.

Bonus points

- Experience shipping software on physical machines or autonomous vehicles in production.

- Exposure to handling data at the petabyte range or coordinating training across many machines at once.

Benefits

Compensation and benefits

- Base salary: $200K to $350K

- Equity

- Health, dental and vision coverage

Location and work model

- San Francisco, CA, United States

- On-site, 6 days in-office per week

- Full-time