Scale-Up Ethernet Engineer
FuriosaAI · Seoul, South Korea · Hwaseong(Dongtan), South Korea
Infrastructureonsite
At a glance
Highlights
- Build Ethernet-based Scale-Up system that links many AI accelerators as one.
- Validate full rack from link bring-up to LLM inference over remote memory.
- Use AI agents to automate bring-up, long tests, log analysis and docs.
Ideal candidate
Systems-minded engineer who sees the rack as one computer and enjoys chasing issues across hardware and software. Comfortable with AI agents and collaborating across teams to make large-scale systems reliable.
Requirements
- Bachelor's in computer, electrical engineering or related field or equivalent practical experience
- Understand data center Ethernet including switch, lossless and RDMA and configure/debug network devices on Linux
- Experience developing or debugging semiconductor system software at register and memory-map level
- Experience tracing issues with hardware, firmware and software engineers
- Experience with rack/server management via BMC/OpenBMC, Redfish, PLDM/MCTP, OpenRMC or rack-manager fleet systems
- Experience building hardware test lab, CI regression, telemetry or long-term stability test infrastructure
- Experience building workflows delegating work to multiple AI agents
- Experience with OCP rack hardware including Open Rack ORv3/ORW, 48V/800VDC, Power Shelf/BBU or rack bring-up with ODM
- Understanding of Scale-Up standards including OCP ESUN/SUE-T, UEC LLR/CBFC, UALink or OCP/SONiC community experience
Lab profile
About FuriosaAI
Builds high-performance, high-efficiency AI compute for the Inference Era.
What they build
- AI inference accelerators
- High-efficiency AI compute hardware
Known for
- Energy-efficient AI chip architecture
- High-performance AI inference
- AI compiler development
Key facts
- Founded in 2017
- Offices in Korea and Silicon Valley
- Compiler-focused R&D lab in Lisbon
Posted · Last seen