Dfx Methodology Architect

AMD AMD · Semiconductors · San Jose, CA · Engineering

This role is for a DFx Methodology Architect at AMD, focusing on defining and improving design-for-testability, debug, characterization, repair, and yield methodologies for complex System-on-Chips (SoCs). The position involves developing new IPs and methodologies, working with functional IP teams, and requires expertise in RTL design, timing constraints, and various design qualification aspects. While the company mentions AI and its role in shaping the future, this specific position is centered on traditional hardware design and verification methodologies, not direct AI/ML model development or deployment.

What you'd actually do

  1. Define DFx Architecture for AMD’s next generation monolithic and stacked SoC product families, including testability, debug, characterization, repair and yield
  2. Address and improve efficiency of timing, PnR, DRC and integration methodologies for DFx IPs
  3. Develop new IPs and methodologies, including process characterization IP (timing, defectivity etc.) for test-vehicles
  4. Define constraints and dependencies for IPs based on block interfaces, power supply & configuration requirements
  5. Work with functional IP teams on integration, analysis and qualification methodologies for a growing number of DFx IPs, tools and flows

Skills

Required

  • RTL design
  • timing constraints
  • design methodologies
  • STA
  • RTL-DRC
  • CDC
  • RDC
  • Constraints checking
  • Perl
  • TCL
  • Python
  • Verilog simulation
  • coverage analysis
  • assertions

Nice to have

  • Scan design
  • ATPG
  • Memory BIST
  • Repair
  • harvesting for yield