System Software Architect

NVIDIA's Performance Lab (PerfLab) builds automation platforms that evaluate the performance and quality of accelerated computing workloads efficiently and accurately. We are looking for a System Software Architect to set the technical direction for these platforms.

Responsibilities:

• Define the technical direction and architecture for major areas of PerfLab's benchmark infrastructure
• Lead the implementation of reusable software, services, and workflows that automate benchmark definition, execution, result collection, validation, and reporting
• Establish engineering approaches that improve the reliability, scalability, observability, maintainability, and reproducibility of large benchmark campaigns
• Stay hands-on with AI benchmarking, analysis, and benchmark design
• Lead delivery across team boundaries and align partner teams on direction
• Raise the engineering bar through architecture and code reviews, technical standards, and mentoring
• Contribute to the strategy and development of internal and open-source projects

Requirements:

• M.Sc. or Ph.D. in Computer Science, Electrical or Computer Engineering
• 10+ years of hands-on experience in developing system software, distributed systems, or AI systems
• Strong Python and software engineering expertise
• Proven technical leadership on sophisticated, multi-functional projects
• Experience in quantitative analysis of AI hardware and systems
• Excellent communication, influence, presentation, and cross-functional coordination skills

Nice to Have:

• Good technical understanding of modern compute hardware
• Working knowledge of how AI workloads stress hardware
• Practical experience building custom agent harnesses or strategies
• Hands-on experience with AI training or inference frameworks
• Demonstrated success turning loosely defined, cross-organizational problems into durable platforms