Senior GPU Kernel Engineer - High-Performance Inference

CoreWeave; Sunnyvale (CA); On-site; USD 182,000 - 242,000; Full time

Benefits offered by this job

Healthcare coverage

Equity awards

401(k) match

Flexible PTO

Tuition reimbursement

ESPP

Parental leave

Lunch provided

Job summary

CoreWeave, the Essential Cloud for AI, is seeking a Senior Engineer for the Benchmarking & Performance team in Sunnyvale, CA. You will write, profile, and tune CUDA kernels on the critical path of large-scale model serving to maximize throughput and minimize latency.

You will lead end-to-end benchmarking efforts including MLPerf submissions, collaborate with product, orchestration, and hardware teams, and mentor junior engineers while raising coding and testing standards.

Qualifications

• 5+ years of experience in HPC, GPU software, or performance-critical systems.
• Hands-on CUDA experience with custom kernel development and optimization.
• Deep understanding of GPU architecture, tensor cores, memory hierarchy, and profiling.
• Strong coding in C++ and Python; reading/writing low-level performance-sensitive code.
• Familiarity with model-serving stacks (vLLM, TensorRT-LLM, llm-d, SGLang).

Responsibilities

• Author, profile, and optimize CUDA kernels on the LLM inference path.
• Tune occupancy, memory coalescing, and data movement for max throughput and min latency.
• Develop benchmarking workflows and reproduce results for MLPerf inferences and training.
• Lead design reviews, decompose multi-service work into milestones, and raise standards.
• Mentor junior engineers and ensure reproducible benchmarking practices.

Skills

CUDA

C++

Python

GPU-architecture

Nsight Compute

Education

Bachelor's degree in CS/Math/Engineering

Tools

CUDA toolkit

Nsight Compute

TensorRT

Job description

CoreWeave, the Essential Cloud for AI, is seeking a Senior Engineer for the Benchmarking & Performance team in Sunnyvale, CA. You will write, profile, and tune CUDA kernels on the critical path of large-scale model serving to maximize throughput and minimize latency.

You will lead end-to-end benchmarking efforts including MLPerf submissions, collaborate with product, orchestration, and hardware teams, and mentor junior engineers while raising coding and testing standards.