About this opportunity
Confidential lists this GPU Kernel Developer for AI Model Evaluation opportunity in Cokeburg, Pennsylvania. Review the employer’s description below for duties, qualifications and application requirements.
Job description
Role Overview
Evaluate GPU and accelerator kernel-development tasks that support training and evaluation of advanced AI models. You will assess numerical correctness, benchmark fairness, task scope, and compilation and runtime validity across a range of kernel task types, then deliver clear written feedback using defined rubrics.
Key Responsibilities
Review the quality, correctness, and completeness of GPU and accelerator kernel-development tasks.
Evaluate numerical accuracy, including appropriate absolute, relative, and ULP tolerances and reference-implementation choices.
Assess performance benchmarking practices for fairness and validity.
Identify issues involving compilation, runtime behavior, task scoping, and kernel implementation quality.
Provide clear, rubric-based written feedback.
Qualifications
3+ years of hands-on experience developing, optimizing, or verifying GPU or accelerator kernels in at least two of CUDA, Triton, NKI, or Pallas for JAX.
Strong understanding of kernel numerical-correctness criteria, including absolute, relative, and ULP tolerances and reference-implementation selection.
Experience profiling and benchmarking performance with tools such as Nsight, NCU, roofline analysis, or framework-native profilers.
Familiarity with common compilation and runtime failures, including driver mismatches, out-of-memory errors, launch-configuration errors, shape or stride mismatches, and autotuning failures.
Experience with at least three kernel task types: generation from specification, translation or lowering across frameworks, migration between hardware targets, debugging, performance optimization, or operator fusion.
Preferred Qualifications
Experience in both NVIDIA GPU ecosystems, including CUDA or Triton, and custom accelerator ecosystems, including NKI, Pallas, or TPU.
Background in compiler engineering, MLIR, or intermediate-representation lowering.
Knowledge of memory-hierarchy optimization, including shared-memory tiling, register pressure, bank conflicts, and memory-coalescing patterns.
Contributions to kernel libraries such as cuBLAS, cuDNN, Triton community kernels, or JAX/XLA custom calls.
Work Terms
Remote, United States.
Hourly engagement.
Compensation
$70 to $90 per hour.
Who can apply
Review the original listing for work authorization, qualifications and employer requirements.