AI Networking
Core
Design and scale high-performance networking fabrics (Ethernet/InfiniBand) connecting hundreds of thousands of GPUs for frontier AI training clusters and Copilot systems.
Role type
Senior IC AI Networking Engineer
Builds
Distributed Ethernet and InfiniBand fabrics for hyperscale AI training clusters
Domain
AI Infrastructure / High-Performance Computing / Datacenter Networking
Deliverable
production ML models
Required skills
ROCE transport design, congestion control, network modeling, telemetry, observability, routing techniques, performance benchmarking, root-cause analysis
Preferred skills
Network co-design, automated troubleshooting, pretraining compute roadmap development
Technologies
ROCE, InfiniBand, Ethernet, ECN, WRED, DCTCP
Responsibilities
Design advanced ROCE transport and congestion control mechanisms; Plan fabric architecture and topology; Develop telemetry and observability solutions; Tune novel routing techniques for reliability; Benchmark AI training and inference clusters; Analyze root causes of performance issues
Seniority
Senior, hands-on IC