Senior ASIC Engineer

NVIDIA 

📍 Shanghai, China 🇨🇳

full-time
senior
Expired
Posted —
This job posting has expired View All ASIC Design Engineer Jobs

Key Skills

ASICRISC-VRTLSynthesisSoC

Industry

SemiconductorConsumer Electronics

Job Description

We are looking for a system-level ASIC design engineer for the PMU and SECIP team. Our team owns three critical GPU IPs — PMU (GPU Power Management), GSP (GPU Service Processor), and SEC (GPU Security Engine) — each of which instantiates a RISC-V-based subsystem called Peregrine, developed by a neighboring team. As GPU systems grow increasingly complex and software teams continue to push more workloads onto these engines, we need a system-level ASIC engineer to help gather Peregrine configuration requirements for all three engines and drive the integration of the Peregrine IP into each of them.

What You'll Be Doing

  • Analyze architectural requirements to define the Peregrine configuration for PMU, GSP, and SEC engines, including how each integrates into the GPU chip.
  • Develop essential wrapper logic as needed to bridge the Peregrine subsystem with each engine's interfaces.
  • Drive the SOCD effort to integrate Peregrine into GPU engines, from RTL through synthesis handoff.
  • Collaborate with the Physical Design team on partition assignment and floorplan considerations.
  • Work with software teams to evaluate system-level performance — particularly memory access latency and bandwidth — ensuring it meets firmware requirements.

What We Need To See

  • BS or MS in Electrical Engineering or Computer Engineering. MS preferred.
  • 3+ years of relevant work experience.
  • Strong communication skills, as this role requires frequent collaboration across multiple teams (architecture, software, physical design, and IP).
  • Solid foundation in front-end ASIC design (RTL coding, synthesis, lint/CDC).
  • Self-motivated with excellent analytical and problem-solving skills.

Ways To Stand Out From The Crowd

  • Experience in system-level chip integration or SoC-level design.
  • Familiarity with RISC-V processor subsystems or embedded firmware workflows.
  • Prior exposure to multi-die / chiplet architectures.

, , JR2015848