Senior Performance Instrumentation Engineer (CUDA Telemetry)

Thomas To; Santa Clara (CA); Sur place; USD 168 000 - 270 000; Plein temps

Avantages offerts par ce poste

Equity

Résumé du poste

NVIDIA is looking for a Software Engineer, Performance Instrumentation to join our Santa Clara team. You will define data requirements, close visibility gaps, and implement low-overhead instrumentation across ML workloads and system software.

You’ll collaborate with partner teams to ensure data quality, measure overhead, and deliver robust telemetry for high-performance computing. The role requires a strong background in Python and C++, CUDA expertise, and experience with PyTorch or JAX.

Qualifications

• Bachelor’s degree in Computer Science, Computer Engineering, or a related field, or equivalent experience.
• 8+ years of relevant software engineering experience in instrumentation or tracing in production ML or systems software.
• Proficiency in CUDA, and working knowledge of PyTorch or JAX, plus NVTX annotation/tracing APIs.
• Experience maintaining data quality in a pipeline with validation, checks, alerting, and provenance.
• Ability to measure instrumentation overhead and trade-offs among overhead, fidelity, sampling, and aggregation.
• Track record implementing changes in code managed by other groups, with clear interfaces and delivery.
• Strong debugging across component boundaries with code reviews, tests, and automated tools.
• Excellent communication skills to convey evidence, uncertainty, and next steps to engineers and data users.
• Proficiency in Python and C++.

Responsabilités

• Define and maintain performance data requirements from profiling libraries, annotation APIs, frameworks, runtimes, and telemetry.
• Identify visibility gaps and collaborate with teams to address or build solutions.
• Inform build of low-overhead instrumentation and annotation APIs for consumers.
• Ensure data quality at ingestion with validation, sanitization, checks, and alerting.
• Measure overhead and observer effects on real workloads to keep within acceptable limits.
• Collaborate with multiple partner teams, managing commitments and progress.

Connaissances

Python

C++

CUDA

Annotation/Tracing APIs (NVTX)

Formation

Bachelor’s degree in CS/EE or equivalent experience

Outils

NVTX

Profiling libraries

Annotation APIs

Description du poste

NVIDIA is looking for a Software Engineer, Performance Instrumentation to join our Santa Clara team. You will define data requirements, close visibility gaps, and implement low-overhead instrumentation across ML workloads and system software.

You’ll collaborate with partner teams to ensure data quality, measure overhead, and deliver robust telemetry for high-performance computing. The role requires a strong background in Python and C++, CUDA expertise, and experience with PyTorch or JAX.