Nvidia 4 дня назад

Senior Manager, Host and Systems Performance

  • Israel, Yokneam
  • Опыт от 10 лет
  • Полная занятость
AINetworkingEthernetSystems EngineeringHPCInfiniBandRoCERDMAPerformance AnalysisDOCANCCLMPI

Описание

Responsibilities

  • Lead, coach, and develop a team focused on NIC, host, DOCA networking stack, and system-level performance.
  • Define performance test plans, methodologies, metrics, and success criteria for new NVIDIA networking technologies.
  • Guide pre-silicon and post-silicon performance planning, analysis, reporting, and readiness reviews.
  • Benchmark and profile workloads across RDMA, RoCE, InfiniBand, Ethernet, MPI, NCCL, DOCA, storage, security, and host networking stacks.
  • Identify bottlenecks across NIC, DPU, CPU, memory, PCIe, firmware, drivers, Linux networking, DOCA, and full-system architecture.
  • Lead root-cause analysis, coordinate mitigation plans, and help hardware, firmware, software, architecture, validation, and product teams align on performance goals.

Requirements

  • B.Sc. or M.Sc. in Computer Science, Computer Engineering, Electrical Engineering, or equivalent experience.
  • Experience in performance analysis, systems engineering, networking, or HPC/AI infrastructure.
  • 10+ years of software engineering experience, including 4+ years in a leadership or management role.
  • Experience with high-performance networking technologies such as RDMA, RoCE, InfiniBand, Ethernet, MPI or NCCL.
  • Hands-on experience with system performance analysis, benchmarking, profiling, and root-cause analysis.
  • Understanding of host architecture, including CPUs, memory hierarchy, NUMA, PCIe, Linux OS, drivers, firmware, and DPU/NIC interactions.
  • Experience creating performance test plans for pre-silicon or post-silicon phases.
  • Programming or scripting experience with Python and Bash; C/C++ experience is helpful.
  • Ability to communicate technical findings clearly and work across engineering teams.

Nice to have

  • Experience with NICs or DPU architecture, networking offloads, host datapath behavior, networking services, or DPU offloads.
  • Experience with AI/HPC cluster performance, distributed training, inference workloads, collective communication libraries, telemetry pipelines, benchmark automation, CUDA, NCCL internals.
  • Linux kernel networking, DPDK, OVS, storage acceleration, or security offloads can also help you succeed in this role.

We offer

  • We support our people with competitive compensation, health and wellness programs, financial benefits, time away from work, family support, learning opportunities, and resources that help employees do their best work.

Hiring process

  • Click here to apply now!