GPU Kernel Expert

Cashback24x7; Indiana; Hybride; USD 96 000 - 124 000; Temps partiel

Avantages offerts par ce poste

Referral reward

Remote work

Résumé du poste

Mercor is seeking a GPU Kernel Expert for an hourly remote contract based in the United States. You will evaluate and improve GPU/accelerator kernel tasks, ensuring numerical correctness and performance.

Responsibilities include identifying failure modes, benchmarking, and contributing to kernel-related projects in a frontier AI lab environment. The role requires 3+ years of hands-on kernel work with CUDA/Triton/NKI/Pallas or JAX, plus strong profiling and optimization skills.

Qualifications

• 3+ years of hands-on experience developing, optimizing, or verifying GPU/accelerator kernels in at least two of: CUDA, Triton, NKI, or Pallas (JAX)
• Strong understanding of numerical-correctness criteria for kernels (absolute/relative/ULP tolerances, reference-implementation selection)
• Demonstrated experience with performance profiling and benchmarking (nsight, ncu, roofline analysis, or framework-native profilers)
• Familiarity with common compilation and runtime failure modes (driver mismatches, OOM, launch-configuration errors, shape/stride mismatches, autotuning failures)
• Experience with at least three kernel task types: generation from specification, translation/lowering across frameworks, migration between hardware targets, debugging, performance optimization, or operator fusion

Connaissances

CUDA

Triton

NKI

Pallas

JAX

Nsight

ncu

Roofline

OOM

Outils

MLIR

IR lowering

Description du poste

GPU Kernel Expert

Company: Mercor

Compensation: $70-$90 per hour

Commitment: Hourly contract

Location: Remote — United States

Referral Reward: Earn up to $1,440 per successful referral

Referral Link: https://t.mercor.com/W0hg1

Location requirements: United States

Evaluate the quality, correctness, and completeness of GPU/accelerator kernel development tasks used to train and evaluate a frontier AI lab's models. You'll assess numerical correctness, performance-benchmarking fairness, task scoping, and compilation/runtime validity across diverse kernel task types — and provide clear, rubric-based written feedback.

Basic Qualifications

• 3+ years of hands-on experience developing, optimizing, or verifying GPU/accelerator kernels in at least two of: CUDA, Triton, NKI, or Pallas (JAX)
• Strong understanding of numerical-correctness criteria for kernels (absolute/relative/ULP tolerances, reference-implementation selection)
• Demonstrated experience with performance profiling and benchmarking (nsight, ncu, roofline analysis, or framework-native profilers)
• Familiarity with common compilation and runtime failure modes (driver mismatches, OOM, launch-configuration errors, shape/stride mismatches, autotuning failures)
• Experience with at least three kernel task types: generation from specification, translation/lowering across frameworks, migration between hardware targets, debugging, performance optimization, or operator fusion

Preferred Qualifications

• Experience across both NVIDIA GPU (CUDA/Triton) and custom-accelerator (NKI/Pallas/TPU) ecosystems
• Background in compiler engineering, MLIR, or intermediate-representation lowering
• Understanding of memory-hierarchy optimization (shared-memory tiling, register pressure, bank conflicts, coalescing patterns)
• Contributions to kernel libraries (cuBLAS, cuDNN, Triton community kernels, JAX/XLA custom calls)

We consider all qualified applicants without regard to legally protected characteristics and provide reasonable accommodations upon request.

Contract and Payment Terms

You will be engaged as an independent contractor. This is a fully remote role that can be completed on your own schedule. Projects can be extended, shortened, or concluded early depending on needs and performance. Your work at Mercor will not involve access to confidential or proprietary information from any employer, client, or institution. Payments are weekly on Stripe or Wise based on services rendered. Please note: We are unable to support H1-B or STEM OPT candidates at this time.

About Mercor

Mercor partners with leading AI labs and enterprises to train frontier models using human expertise. You will work on projects that focus on training and enhancing AI systems. You will be paid competitively, collaborate with leading researchers, and help shape the next generation of AI systems in your area of expertise.