GPU Kernel Expert

Remote | Independent Contractor | United States | $145,600–$187,200 annualized ($70–$90/hour)

About the Role

Apply your GPU kernel development and optimization expertise to help improve the capabilities of advanced AI systems.

As a GPU Kernel Expert, you’ll evaluate GPU and accelerator kernel development tasks used to train and assess frontier AI models. You’ll assess numerical correctness, performance benchmarking, task scope, compilation, and runtime behavior across a range of kernel development scenarios.

Your technical judgment will help ensure that AI training and evaluation tasks accurately reflect real-world GPU and accelerator engineering challenges.

What You’ll Do

• Evaluate GPU and accelerator kernel tasks for quality, correctness, completeness, and technical rigor.

• Assess numerical correctness using appropriate absolute, relative, and ULP tolerance criteria.

• Evaluate whether reference implementations are appropriate and reliable for validating kernel outputs.

• Review performance benchmarks for fairness, consistency, and methodological soundness.

• Assess task scope and determine whether requirements are technically clear, realistic, and appropriately defined.

• Evaluate compilation and runtime behavior across different hardware and software environments.

• Identify common failure modes, including:

  • Driver and environment mismatches

  • Out-of-memory (OOM) conditions

  • Kernel launch configuration errors

  • Shape and stride mismatches

  • Autotuning failures

• Provide clear, structured, rubric-based written feedback.

• Evaluate kernel tasks involving generation, translation, migration, debugging, optimization, and operator fusion.

What You Bring

• 3+ years of hands-on experience developing, optimizing, or verifying GPU or accelerator kernels.

• Professional experience with at least two of the following:

  • CUDA

  • Triton

  • NKI

  • Pallas (JAX)

• Strong understanding of numerical correctness for GPU kernels, including:

  • Absolute and relative tolerances

  • ULP-based comparisons

  • Reference implementation selection

• Demonstrated experience with performance profiling and benchmarking using tools or methodologies such as:

  • NVIDIA Nsight

  • Nsight Compute (NCU)

  • Roofline analysis

  • Framework-native profilers

• Familiarity with common GPU/kernel compilation and runtime failure modes.

• Experience with at least three of the following kernel task types:

  • Generation from specification

  • Translation or lowering across frameworks

  • Migration between hardware targets

  • Kernel debugging

  • Performance optimization

  • Operator fusion

• Strong analytical skills and attention to technical detail.

• Ability to communicate complex kernel engineering concepts clearly in writing.

Preferred Qualifications

Experience in the following areas is highly valuable:

• Experience across both NVIDIA GPU ecosystems such as CUDA/Triton and custom accelerator ecosystems such as NKI/Pallas/TPU.

• Background in compiler engineering, MLIR, or intermediate-representation lowering.

• Deep understanding of memory-hierarchy optimization, including:

  • Shared-memory tiling

  • Register pressure

  • Bank conflicts

  • Memory coalescing

• Contributions to GPU or accelerator kernel libraries and related open-source projects.

• Experience with technologies such as cuBLAS, cuDNN, Triton community kernels, or JAX/XLA custom calls.

Compensation & Engagement

• Rate: $70–$90/hour

• Annualized Equivalent: $145,600–$187,200

• Location: United States

• Work Arrangement: Fully remote

• Engagement Type: Independent contractor

• Schedule: Flexible

• Payment: Weekly via Stripe or Wise

Annualized compensation is based on 2,080 hours per year for comparison purposes only. Actual earnings depend on the number of hours and projects completed.

Why This Opportunity?

• Apply your GPU and accelerator engineering expertise to advanced AI development.

• Work on technically challenging kernel evaluation and optimization problems.

• Help improve how AI systems understand low-level performance, numerical correctness, and hardware acceleration.

• Evaluate realistic engineering scenarios across GPUs, accelerators, frameworks, and compilers.

• Work remotely with a flexible schedule.

• Contribute specialized expertise to high-impact AI training and evaluation projects.

Contract & Payment Terms

• You will be engaged as an independent contractor.

• Work is fully remote and can be completed on your own schedule.

• Projects may be extended, shortened, or concluded early depending on project needs and performance.

• Your work will not require access to confidential or proprietary information belonging to any employer, client, or institution.

• Payments are made weekly via Stripe or Wise based on services rendered.

• H-1B and STEM OPT candidates cannot be supported at this time.

Equal Opportunity

All qualified applicants will be considered without regard to legally protected characteristics. Reasonable accommodations are available upon request.