Director, Rack Scale Software Architecture

NVIDIA is defining the next era of computing, where the GPU acts as the brain of computers, robots, and self-driving cars. They are looking for a Director to lead the software architecture for Rack-Scale Systems. As AI models grow, single GPUs aren't enough; companies need entire racks of HGX/DGX systems connected via NVLink to function as one giant computer. Your job is to lead the team building the software stack (Firmware, Kernel, OS, Networking) that makes this "Scale-Up" architecture possible.

Key Responsibilities

• End-to-End Architecture: Drive the software architecture for NVIDIA's rack-scale products, ensuring high reliability and performance at the SW/HW interface.
• Strategic Roadmap: Translate forward-looking plans into clear software requirements that anchor execution across the organization.
• Hyperscaler Engagement: Work directly with major cloud providers (AWS, Azure, Google, Meta) to align their roadmaps with NVIDIA’s vision.
• Technical Leadership: Lead the team responsible for Firmware, Kernel Drivers, OS, Networking, and Fabrics. You will make key technical decisions amidst ambiguity.
• Team Management: Provide career mentorship and technical guidance to a high-performing engineering team.

Requirements

• Experience: 15+ years of system architecture experience with 8+ years in management.
• Technical Domain: Deep experience in designing architecture for scalable server systems, specifically at the SW/HW interface.
• Accelerator Knowledge: Previous experience with complex system software for GPUs, DPUs, or FPGAs.
• Leadership: Proven ability to manage large-scale sophisticated code bases and operate in highly matrixed organizations.
• Education: BS or MS in Computer Engineering/Science.

Nice to Have

• Advanced Networking: Strong understanding of Ethernet, Infiniband, CXL, and UCIE architectures.
• Cluster Management: Knowledge of large-scale cloud and cluster-level deployment systems.

Compensation

• Base Salary: $320k - $488k USD.
• Equity: NVIDIA RSUs (Restricted Stock Units) are a significant component of compensation, likely doubling the total package value given the stock's performance.