Software Development Engineer - GPU Interconnects and Ethernet AI
Core
Design, develop, and optimize networking software for GPU scale-up interconnects and Ethernet-based AI fabrics on AMD Helios rack-scale systems.
Role type
Senior IC systems software engineer (AI networking)
Builds
High-performance, low-latency networking for next-generation AI infrastructure
Domain
Datacenter networking, AI infrastructure, GPU interconnects
Deliverable
production ML models
Required skills
Linux kernel driver development, embedded firmware development, C/C++ programming, Ethernet networking (MAC, PCS, FEC, SerDes), GPU interconnect technologies, switch software (SONiC, SAI), high-performance transport protocols (RDMA, RoCE, InfiniBand)
Preferred skills
Python or Go for tooling, performance analysis and profiling, debugging complex multi-node system issues, RAS and serviceability considerations
Technologies
Linux kernel, PCIe, DMA, RTOS, SONiC, SAI, RDMA, RoCE, InfiniBand, UALink, Infinity Fabric, NVLink
Responsibilities
Lead design and development of networking software for GPU scale-up interconnect and Ethernet-based AI fabrics; Develop and optimize Linux kernel drivers, embedded firmware, and user-space software; Drive hardware and software bring-up of new ASICs and platforms; Work with switch software teams on configuration, forwarding, QoS, and fabric management; Define firmware, driver, and management interfaces; Analyze end-to-end network performance and latency; Develop tools and telemetry frameworks for monitoring link and fabric health; Investigate and resolve issues across the stack; Contribute to resiliency and deployment readiness features; Support performance characterization of collective operations; Partner with internal and external teams to validate interoperability; Provide technical leadership through design reviews and mentorship
Seniority
Senior, hands-on IC
