Senior Software Engineer – Networking

Tech Stack

Solid badges = required, outlined = preferred

Responsibilities

• Build networking systems for running models across multiple Mother systems, supporting intra-node tensor parallelism and inter-node pipeline parallelism.
• Design and optimize high-throughput, low-latency activation transfer between systems while keeping weight-heavy execution local.
• Work with RDMA/RoCEv2, GPUDirect RDMA, NVIDIA ConnectX-7-class NICs, dual-port 200 Gbps links, and PCIe Gen 5 x16 bandwidth.
• Measure and improve throughput, latency, jitter, CPU overhead, memory-copy behavior, congestion, and failure recovery.
• Build observability and diagnostics for NIC behavior, RDMA fabric health, GPU memory movement, and multi-node inference performance across the stack from Linux to inference runtime.

Soft Skills

Benefits

• Dental
• Health Insurance
• Life Insurance
• Unlimited PTO
• Vision

Culture

Requirements