Supercomputing Engineer
San Francisco Compute CompanySan Francisco, California, United States
Posted October 31, 2025
170000 - 300000 USD year
Full Time
About the Company
San Francisco Compute Company aims to revolutionize the compute market by creating a real-time trading platform for compute contracts. This venue allows startups to access flexible compute resources without long-term commitments, while ensuring providers maintain profitability through speculative trading. By streamlining the supply chain and matching buyers with sellers efficiently, SF Compute seeks to make compute as accessible and reliable as other commodities.
Job Description
Compute is a commodity provider specializing in real-time compute trading, similar to how oil and hardware are commodities. They are building a trading venue for large compute contracts, aiming to reduce friction and bring a new type of participant into the supply chain. SF Compute is a small, data-driven team focused on scaling their compute clusters and maintaining efficient operation.
Requirements
- Managed at least one GPU training cluster.
- Deep understanding of Linux, networking fundamentals, CUDA, NCCL, and Infiniband.
- Automation of hardware deployments using IaC
- Strong understanding of Kubernetes implementation
- Experience with HPC network architectures (BGP, fat-tree, VXLAN, MCLAG, etc.)
Benefits
- Competitive salary
- Equity
- Team members offered equity
- Unlimited paid time off
- Medical, dental & vision insurance
- Unlimited office book budget
- Unlimited paid time off, with biological, adoptive, and foster parents paid time off
- Daily lunch
- Unlimited paid time off
- Parental leave
- 300k - 400k compensation
