Network Engineer, Supercomputing
Core
Own the lowest layers of the network stack for large-scale AI training and inference, ensuring interconnect reliability across GPU fabrics.
Role type
Senior IC network engineer (supercomputing/AI infrastructure)
Builds
Reliable RDMA/RoCE and NVLink/NVSwitch fabrics for distributed GPU clusters
Domain
AI/ML infrastructure, high-performance computing, cloud networking
Deliverable
production ML models
Required skills
Linux debugging, Python or Rust, Kubernetes/Slurm, RDMA/RoCEv2, NVLink/NVSwitch, cluster operations
Preferred skills
Host-level Linux tooling, multi-cloud network primitives, CUDA/NCCL profiling, statistical reliability reasoning, deep learning framework architecture
Technologies
RDMA, RoCEv2, NVLink, NVSwitch, NCCL, Kubernetes, Slurm, Python, Rust, Linux
Responsibilities
Debug production collectives down to the NIC, build network instrumentation and dashboards, drive escalations with cloud-provider networking teams, validate GPU network fabric design, diagnose collective failures and congestion control behavior
Seniority
Senior, hands-on IC