Senior Technical Program Manager - AI Cloud Benchmarking

NVIDIA Corporation; Santa Clara (CA); On-site; USD 168,000 - 322,000; Full time

Job summary

NVIDIA Corporation is seeking a Senior Technical Program Manager to run the Exemplar Performance Program, guiding CSPs, NVIDIA Cloud Partners, and internal teams. The role focuses on capacity, validation, debugging, and release planning across training and inference benchmarks.

You will coordinate with engineering, product, and partner-facing groups, manage cross-functional workstreams, and scale as the program expands into new workloads and stack complexities.

Qualifications

• 8+ years of technical program management experience.
• Deep understanding of cloud infrastructure, inference systems, and distributed computing.
• Experience coordinating multi-stakeholder programs across time zones and organizations.
• Experience with software release management, QA coordination, and platform validation cycles.

Responsibilities

• Maintain demand and allocation levels across platforms, GPUs count, topologies, and workloads.
• Coordinate capacity reservations, readiness checks, and debug windows with cluster owners.
• Track cross-CSP performance investigations and build pre-flight checks and playbooks.
• Lead release readiness summaries for each program increment and ensure alignment with roadmaps.
• Develop metrics for program efficiency and communicate progress to senior leadership.

Skills

Technical program management

Cloud platforms

Distributed computing

Education

Bachelor's degree in Computer Science, Electrical Engineering, AI, or related field

Tools

Jira

NVBugs

Job description

NVIDIA Corporation is seeking a Senior Technical Program Manager to run the Exemplar Performance Program, guiding CSPs, NVIDIA Cloud Partners, and internal teams. The role focuses on capacity, validation, debugging, and release planning across training and inference benchmarks.

You will coordinate with engineering, product, and partner-facing groups, manage cross-functional workstreams, and scale as the program expands into new workloads and stack complexities.