Senior Power and Performance Architect

NVIDIA NVIDIA · Semiconductors · Tel Aviv, Israel

NVIDIA is seeking a Senior Power and Performance Architect to optimize network systems for AI clusters. The role involves developing and validating power and performance optimization techniques for network fabrics, components, and systems, collaborating with cross-functional teams including AI researchers. Requires experience in network architecture, design, performance analysis, and familiarity with AI/ML workload network demands.

What you'd actually do

  1. Develop, model, and validate groundbreaking power and performance optimization techniques for network fabrics, components (NICs, Switches), and systems.
  2. Collaborate closely with cross-functional teams including silicon design, system architects, software/firmware engineers, performance analysts, thermal engineers, and AI researchers to ensure end-to-end power and performance optimization.
  3. You will determine strategies that successfully implement flawless performance and power efficiency in our ambitious projects!

Skills

Required

  • BSc or MSc or equivalent experience in Electrical Engineering, Computer Engineering, Computer Science or a related field
  • 5+ years of relevant experience in network architecture, design, or performance analysis
  • Solid understanding of power consumption dynamics in network hardware (NICs, switches, cables) and systems
  • Proven experience in network architecture and design, particularly for large-scale systems (Data Center, HPC, AI Clusters)
  • Strong understanding of network protocols (Ethernet, InfiniBand, RoCE, TCP/IP) and their impact on performance and efficiency
  • Demonstrable expertise in network performance analysis, bottleneck identification, and tuning
  • Familiarity with the characteristics and network demands of AI/ML workloads
  • Utmost passion for attention to details in design and a high focus on design quality, particularly concerning power/performance trade-offs

Nice to have

  • Advanced degree or equivalent experience in a related field
  • Proven dedication to system-level power/performance trade-off analysis, especially in distributed computing or large-scale network environments
  • Experience in power modeling, measurement techniques, or relevant tools for network components and systems
  • Proficiency with network simulation tools (e.g., ns-3, OMNeT++, proprietary tools) or performance modeling frameworks
  • Understanding of silicon-level power characteristics and optimization techniques

What the JD emphasized

  • network architecture, design, or performance analysis
  • network architecture and design
  • network performance analysis
  • AI/ML workloads
  • power/performance trade-offs