Power Architect, AI Data Center Chiplets

Tenstorrent · Semiconductors · Santa Clara, CA · Architecture

The role focuses on optimizing the energy efficiency of RISC-V based CPUs and AI Data Centers for Tenstorrent, a company at the forefront of AI technology. The Power Architect will be responsible for power management, SoC power architecture, power delivery networks, thermal analysis, and performance trade-offs, with a specific emphasis on analyzing AI and ML workloads for performance and efficiency. This is a hybrid role based in Santa Clara, CA, with opportunities for growth and impact in the AI hardware design space.

What you'd actually do

  1. Deep expertise in Power Management, SoC Power Architecture, Power Delivery Networks, Thermal Analysis, and Performance Trade-offs
  2. Proficiency in Power Estimation Methodologies (e.g., using PTPX inputs) and trade-off modeling
  3. Experience with silicon PVT corners, process node trade-offs, binning, voltage droop mitigation techniques
  4. Ability to analyze AI and ML workloads for performance and efficiency
  5. Strong debugging skills and ability to design experiments to root cause issues

Skills

Required

  • Power Management
  • SoC Power Architecture
  • Power Delivery Networks
  • Thermal Analysis
  • Performance Trade-offs
  • Power Estimation Methodologies
  • Silicon PVT corners
  • Process node trade-offs
  • Binning
  • Voltage droop mitigation techniques
  • Analyze AI and ML workloads
  • Debugging skills
  • Design experiments
  • Root cause analysis
  • Collaboration skills
  • Interpersonal skills

Nice to have

  • RISC-V
  • Chiplets
  • Open Chiplet Architecture
  • OCMI
  • PTPX

What the JD emphasized

  • energy efficiency of RISC-V based CPUs and AI Data Centers
  • Power Management and Optimization of SoCs and Chiplets
  • fast paced technical environment with multi-disciplinary teams
  • AI and ML workloads for performance and efficiency
  • debugging, experimentation, and root cause analysis
  • multidisciplinary teams across silicon, firmware, and systems
  • energy efficiency challenges in data centers

Other signals

  • AI hardware design
  • energy efficiency
  • RISC-V CPUs
  • Data Center Products