Principal AI Network Hardware Systems Engineer
Core
Define and develop networking requirements for large-scale AI training and inference clusters, optimizing RDMA-based solutions and validating performance across distributed AI workloads.
Role type
Principal AI Network Hardware Systems Engineer
Builds
Scalable networking solutions for AI clusters from concept through datacenter deployment
Domain
AI Infrastructure / High-Speed Networking
Deliverable
production ML models | infrastructure
Required skills
RDMA technologies, AI fabrics, distributed training environments, RoCE, congestion control, ECN, PFC, DCQCN, AI/ML workload communication patterns, collective operations, SONiC, Linux networking, network telemetry, network operating systems, network switches, SmartNICs, DPUs, NIC offloads, large-scale cloud infrastructure, AI networking technologies, Ultra Ethernet, hyperscale AI cluster architectures, network stress tools, validation frameworks, performance benchmarks, observability solutions, packet analysis tools, telemetry infrastructure, network automation frameworks, high-speed networking environments (200G/400G/800G Ethernet)
Preferred skills
Experience developing network stress tools, validation frameworks, performance benchmarks, or observability solutions
Technologies
RDMA, RoCE, ECN, PFC, DCQCN, SONiC, Linux, 200G/400G/800G Ethernet
Responsibilities
Define network concepts of operation, serviceability requirements, telemetry requirements, and operational models for AI infrastructure; Analyze transport-layer behavior and performance characteristics across large-scale distributed AI workloads; Evaluate network protocol implementations and debug issues impacting latency, throughput, scalability, and reliability; Design, validate, and optimize RDMA-based networking solutions for AI clusters; Develop validation methodologies for AI traffic patterns and collective communication workloads; Develop and execute networking validation strategies covering functionality, performance, scale, interoperability, resiliency, and reliability
Seniority
Principal, hands-on IC
