Remote Principal Network Engineer

Description:

• The role is for a Principal Network Engineer at SF Compute, focusing on designing and managing networks for large-scale GPU clusters globally.
• Responsibilities include network architecture design, automated fabric provisioning and validation, performance monitoring and optimization, and interconnecting high-performance distributed storage systems.
• The ideal candidate will design a 400GbE spine-leaf network from scratch and automate its deployment across a 20k node cluster.
• In-person work is preferred at the San Francisco office, but remote applicants will also be considered.

Requirements:

• Candidates must have previous experience with technologies related to HPC and GPU networks, including RoCEv2, InfiniBand, eBGP, EVPN/VXLAN, QoS, and ACLs.
• Experience in architecting resilient high-performance networks, including fat-tree and multi-layer spine-leaf topologies, is required.
• Prior experience with network automation using tools such as Ansible, Bash, and Python is necessary.
• Candidates should be comfortable operating and have opinions about networking hardware from Arista, Cisco, Dell, Juniper, OCP, and SONiC.
• Comfort in configuring Next-Generation Firewalls (NGFWs) is essential.
• A strong appreciation for good documentation is required.

Benefits:

• Team members receive a generous equity grant along with a competitive salary.
• The company sponsors visas and work permits for eligible candidates.
• Retirement plans are matched up to 4% for 401(k) contributions.
• Comprehensive medical, dental, and vision insurance is provided, covering 100% of premiums for employees and their dependents.
• Employees enjoy unlimited paid time off and over 10 observed holidays.
• Paid parental leave is offered for biological, adoptive, and foster parents.
• Daily lunch is covered for employees.
• There is an unlimited office book budget for purchasing books for the office.