NVIDIA
Principal Hardware Security Architect, Advanced Technologies and Threat Modeling
$272K–$431K
US, CA, Santa Clara +1 · Principal · Full-time · Posted Oct 13, 2026
- Location
- US, CA, Santa Clara
US, CA, Remote - Experience
- 15+ years of meaningful relevant work or research experience.
- Education
- MS or PhD (or equivalent experience) in Electrical Engineering, Computer Science, or Computer Engineering.
About the job
Defines security architecture and roadmaps for NVIDIA SoCs and GPUs, with a focus on physical attacks, advanced packaging, photonics, and AI-assisted vulnerability discovery.
What you'll do
- Develop platform security roadmaps and mitigations for physical attacks, including fault injection and side-channel attacks.
- Threat-model advanced packaging and photonics technologies, and guide adoption of security standards such as OCP S.A.F.E.
- Review architecture and verification specifications and evaluate AI-assisted security verification and vulnerability discovery methods.
What they ask for
- Experience with ASIC design processes and large SoC or GPU architecture.
- Programming and debugging experience with languages such as C/C++, Verilog, ARM assembly, and Python/Perl scripting.
- Deep knowledge of SoC security architecture, secure/measured boot, and ARM or RISCV CPU architectures.
- Direct experience mitigating physical attacks; knowledge of threat modeling and penetration testing.
Nice to have
- Research or hands-on experience securing advanced packaging, chiplets, or SiP designs; or expertise in silicon photonics security.
- Experience applying AI/ML to hardware vulnerability discovery, automated triage, or autonomous fuzzing; participation in standards bodies or security patents.
Pay
272,000 USD - 431,250 USD (base salary)
Tools
ASIC, SoCs, GPUs, C/C++, Verilog, ARM assembly, Python/Perl, OCP S.A.F.E.
AI focus
Use AI to discover and triage hardware vulnerabilities, evaluate AI-assisted hardware security verification, and explore autonomous research agents for fuzzing GPU kernel drivers.
Model-written summary · the original posting is authoritative