Senior Compute Kernel Architect, GPU Power

Location: US, CA, Santa Clara

Time Type: Full time

Job Description

NVIDIA is seeking a Compute Kernel Performance Architect who can develop, profile, and analyze CUDA workloads with a strong focus on GPU power behavior. In this role, you will create specialized workloads that exercise the GPU’s compute, memory, and I/O subsystems under demanding operating conditions. You will work closely with GPU architects, power architects, silicon validation engineers, and software teams to characterize workload behavior and influence the power architecture of future NVIDIA products. This position sits at the intersection of GPU architecture, high-performance software, and silicon characterization.

What You'll Be Doing:

• Design and develop CUDA kernels and infrastructure that exercise worst-case power behavior across GPU compute, memory, and I/O subsystems.

• Profile workloads to understand the relationship between kernel behavior, hardware utilization, performance, and power consumption.

• Build workloads that generate controlled steady-state and transient power conditions across multiple GPU architectures.

• Partner with GPU architects and silicon teams to identify functional units and workload patterns that require additional characterization.

• Support power-stress methodology from pre-silicon modeling and simulation through post-silicon bring-up and validation.

What We Need to See:

• MS or PhD or equivalent experience in Computer Science, Electrical Engineering, Computer Engineering, or a related field—or equivalent practical experience.

• 5+ years of experience in CUDA programming, GPU kernel development, high-performance computing, or performance architecture.

• Hands-on experience developing and optimizing GPU kernels, including work at the PTX or assembly level.

• Experience with GPU performance-analysis tools such as Nsight Compute, Nsight Systems, nvprof, or equivalent tools.

• Strong understanding of GPU build principles, including streaming multiprocessors, execution pipelines, memory hierarchy, synchronization, occupancy, and power states.

• Excellent analytical, debugging, and communication skills.

• Ability to work effectively across GPU architecture, software, silicon validation, and hardware engineering teams.

Ways to Stand Out from the Crowd:

• Experience crafting GPU power-stress microbenchmarks or test-to-failure workloads.

• Familiarity with Power Delivery Network concepts, including package and board-level behavior, impedance, inductance, decoupling, resonance, voltage droop, and overshoot.

• Understanding of di/dt and how changes in current over time can compose voltage transients.

• Experience with DVFS, AVFS, clock management, power states, or hardware noise-mitigation mechanisms.

• Knowledge of how software workload patterns can interact with system-level power-delivery behavior.

Our team works at the core of NVIDIA’s GPU performance and power stack. We collaborate closely with Compute Architecture, Power Architecture, Silicon Solutions, circuit-design teams, and deep-learning software teams. The workloads, tools, and analysis produced by this team help validate current products and influence the build of upcoming NVIDIA GPUs.

Your base salary will be determined based on your location, experience, and the pay of employees in similar positions. The base salary range is 184,000 USD - 287,500 USD for Level 4, and 224,000 USD - 356,500 USD for Level 5.Applications for this job will be accepted at least until July 26, 2026.

This posting is for an existing vacancy. 

NVIDIA uses AI tools in its recruiting processes.

NVIDIA is committed to fostering an inclusive work environment and proud to be an equal opportunity employer. As we highly value diversity in our current and future employees, we do not discriminate (including in our hiring and promotion practices) on the basis of race, religion, color, national origin, gender, gender expression, sexual orientation, age, marital status, veteran status, disability status or any other characteristic protected by law.