Senior Engineering Manager, Data Plane Systems

Crusoe · Data AI · San Francisco, CA - US · Cloud Engineering

This role is for a Senior Engineering Manager to lead a team responsible for high-performance SDN data planes across hosts and DPUs, focusing on architecture, implementation, and production operation of a hardware-accelerated networking stack for large-scale GPU workloads. The role involves defining roadmaps, integrating DPUs and hardware accelerators, developing Linux kernel networking components, XDP/eBPF, and DPDK, migrating networking functions to hardware offload, leading performance benchmarking, regression prevention, and incident response, and mentoring a team of systems engineers. Collaboration with control-plane teams is also key.

What you'd actually do

  1. Technical Ownership: You will define the roadmap for SDN data plane systems and lead the integration of DPUs (such as NVIDIA BlueField) and hardware accelerators.
  2. Architecture & Optimization: You will oversee the development of Linux kernel networking components, XDP/eBPF data paths, and DPDK-based fast paths while driving the migration of networking functions to hardware offload architectures.
  3. Production Execution: You will lead performance benchmarking, regression prevention, and incident response, ensuring operational excellence within 3-6 month execution cycles.
  4. People Leadership: You will mentor and grow a team of senior and staff-level systems engineers, setting technical standards and fostering a high-performance culture of accountability.
  5. Collaboration: You will partner closely with control-plane teams (OVN/OVS) to optimize throughput and latency for multi-tenant GPU clusters.

Skills

Required

  • 10+ years in high-performance networking or systems engineering
  • 5-7+ years managing senior/staff-level talent
  • Deep knowledge of Linux networking internals
  • kernel architecture
  • Experience with XDP/eBPF
  • AF_XDP
  • DPDK
  • Hands-on experience with DPU integration
  • migrating networking functions to hardware accelerators
  • Strong understanding of low-latency networking
  • performance tuning
  • benchmarking
  • Ability to resolve complex technical challenges in a fast-moving, execution-heavy environment

Nice to have

  • RDMA/RoCE in GPU environments
  • P4 programmable networking
  • open-source contributions to kernel systems

What the JD emphasized

  • high-performance SDN data planes
  • large-scale GPU workloads
  • hardware-accelerated networking stack
  • aggressive timelines
  • moving features from commit to production within days
  • Linux kernel networking components
  • XDP/eBPF data paths
  • DPDK-based fast paths
  • hardware offload architectures
  • performance benchmarking
  • regression prevention
  • incident response
  • 3-6 month execution cycles
  • senior/staff-level talent
  • Deep knowledge of Linux networking internals
  • kernel architecture
  • XDP/eBPF
  • AF_XDP
  • DPDK
  • DPU integration
  • migrating networking functions to hardware accelerators
  • low-latency networking
  • performance tuning
  • benchmarking
  • fast-moving, execution-heavy environment