Senior Inference Engineer: GPU Kernel Optimizations + Equity

Nvidia Corporation; Santa Clara (CA); On-site; USD 184,000 - 287,500; Full time

Benefits offered by this job

Equity

Comprehensive benefits

Job summary

NVIDIA is seeking a Senior Inference Engineer to push GPU kernel optimization for LLM inference across silicon-level measurements. You will help build benchmarking infrastructure, model-level performance tooling, and optimization policies for production deployments.

The role collaborates with compiler, hardware, kernel, and framework teams to surface bottlenecks and deliver measurable gains, in a full-time, on-site capacity in Santa Clara.

Qualifications

• Masters/PhD in CS/CE or equivalent experience.
• 6+ years of relevant industry experience.
• Experience building or directing agentic AI systems.
• Strong Python and C++ skills with solid software fundamentals.
• Hands-on GPU profiling with CUPTI/NSYS/NCU and bottleneck attribution.
• Experience with LLM inference frameworks (TRT-LLM, SGLang, or vLLM).
• Knowledge of CUDA/CUTLASS/Triton and reading PTX/SASS.

Responsibilities

• Drive GPU kernel microbenchmarking across full configuration space with silicon fidelity.
• Perform end-to-end model performance analysis to inform optimization policies.
• Apply AI-driven analysis to diagnose performance gaps and validate findings with silicon measurements.

Skills

Python

C++

Agentic AI systems

GPU profiling

LLM inference frameworks

CUDA/CUTLASS/Triton

PTX/SASS reading

Education

Master's or PhD in CS/CE or equivalent experience

Tools

CUPTI

NSYS

NCU

TRT-LLM

SGLang

vLLM

PTX/SASS analysis tools

Job description

NVIDIA is seeking a Senior Inference Engineer to push GPU kernel optimization for LLM inference across silicon-level measurements. You will help build benchmarking infrastructure, model-level performance tooling, and optimization policies for production deployments.

The role collaborates with compiler, hardware, kernel, and framework teams to surface bottlenecks and deliver measurable gains, in a full-time, on-site capacity in Santa Clara.