Distinguished Resiliency and Safety Architect, GPU Diagnostics

Role summary

We are now seeking a Resiliency and Safety Architect to support the development of GPU (graphical processing unit) diagnostics for Resiliency in the Datacenter and Functional SafetyThe part of overall safety that depends on a system or equipment operating correctly in response to its inputs, including safe management of faults and failures. Read more in Autonomous Vehicles and Robots. In this role, you will be a key member of a team of innovators, challenging the status quo and pushing beyond boundaries. You will have the opportunity to impact the industry's leading GPUs and SoCs powering product lines ranging from the rapidly growing field of artificial intelligence to self-driving cars and robots.

Responsibilities

• Design, develop, and maintain diagnostics software suite to efficiently stress test the Company GPUs and SOCs to identify hardware defects, including defects that cause silent data corruption. These tests will run in large-scale deployments of Datacenter GPUs and Safety SOCs in package/board/rack configurations spanning GPUs, CPUs, and Networking SOCs.
• Address coverage gaps in the Company diagnostic suite flagged by silicon failures on customer workloads or test suites. Enhance diagnostics to improve repeatability of failures detected and optimize test time.
• Tests for GPUs in automotive functional safety contexts should include low-level routines to exercise instruction sets, memory subsystems and interrupt mechanisms, in compliance with ISO 26262Road Vehicles - Functional Safety. Automotive functional safety standard defining ASILAutomotive Safety Integrity Level (A to D) under ISO 26262, derived from severity, exposure, and controllability. Read more levels and the automotive safety lifecycle. Read more and related safety standards. Collaborate with architecture, RTL, and verification teams to ensure safety coverage, correctness, and robustness across GPU generations.
• Study silent data corruption, intermittent faults, and hard-to-reproduce failures in the field, including customer returns (RMAs), to establish root causes, and improve detection by diagnostics
• Support deployment of diagnostics in pre-production qualification environments as well as large-scale production usages.

Requirements

• Master’s or PhD degree in Computer Science, Computer Engineering, Electrical Engineering or closely related degree or equivalent experience.
• At least 15+ years of relevant experience.
• Ability to reason across hardware/software boundaries to debug complex system-level issues
• In-depth understanding of the architecture and micro-architecture of high-performance computing systems. Strong knowledge of hardware failure mechanisms that can result in incorrect computation.
• Proficiency in C/C++, CUDA programming.
• Scripting and automation with Python or similar.
• Understanding of the software development life cycle, from requirements to testing closure and maintenance, including creating customer releases and documentation.
• Excellent interpersonal skills and ability to collaborate with on-site and remote teams.
• Strong debugging and analytical skills.
• Be self-driven and results oriented.

Nice to have

• Familiarity with GPU and SOC Architectures, Machine Learning/Deep Learning concepts
• Understanding factors causing silent data corruption in hardware
• Ability to use high performance libraries and write hand-crafted kernels where necessary to create stress conditions to induce hardware failures.
• Experience in embedded software development.

The Company's invention of the GPU 1999 sparked the growth of the PC gaming market, redefined modern computer graphics, and revolutionized parallel computing. More recently, GPU deep learning ignited modern AI - the next era of computing - with the GPU acting as the brain of computers, robots, and self-driving cars that can perceive and understand the world. Today, we are increasingly known as “the AI computing company”.

Standards

• ISO 26262