Remote GPU Kernel Engineer - Performance & Debugging

Anyone AI Inc.; Argentina; A distancia; ARS 16.740.000 - 23.436.000; A tiempo parcial

Descripción de la vacante

Anyone AI Inc. is seeking experienced GPU Kernel Engineers to review, debug, and evaluate high-performance compute kernels for AI workloads. You will optimize kernels across CUDA, Triton, and related frameworks, ensuring numerical correctness and reproducibility.

The role focuses on kernel implementation and benchmarking, memory hierarchy optimization, and migration across hardware. This is a remote, part-time consulting engagement appealing to experts who enjoy hardware-close optimization.

Formación

• 3+ years of hands-on experience developing, optimizing, or debugging GPU kernels.
• Two or more of CUDA, Triton, NKI/AWS Neuron, or Pallas/JAX experience.
• Strong understanding of GPU performance and memory hierarchy.
• Experience with kernel profiling tools (Nsight, NCU, roofline, or framework profilers).
• Solid understanding of memory bandwidth, compute throughput, occupancy, shared memory, register pressure, and memory coalescing.
• Fluency in floating-point numerical correctness and tolerance thresholds.
• Experience debugging kernel compilation and runtime issues.
• Ability to distinguish software defects, environment problems, and genuine optimization challenges.

Responsabilidades

• Review GPU and accelerator kernel implementations for correctness.
• Compare outputs against reference implementations.
• Evaluate numerical tolerance thresholds.
• Review kernel benchmarks and determine fairness of comparisons.
• Identify performance bottlenecks and optimization opportunities.
• Assess whether performance targets are realistic given hardware limits.
• Review kernel translations and hardware migrations.
• Identify compilation, memory, shape, and runtime issues.
• Provide clear, actionable technical feedback.

Conocimientos

GPU kernel development

Performance optimization

Herramientas

CUDA

Triton

NKI / AWS Neuron

Pallas / JAX

Nsight

NCU

roofline analysis

Descripción del empleo

Anyone AI Inc. is seeking experienced GPU Kernel Engineers to review, debug, and evaluate high-performance compute kernels for AI workloads. You will optimize kernels across CUDA, Triton, and related frameworks, ensuring numerical correctness and reproducibility.

The role focuses on kernel implementation and benchmarking, memory hierarchy optimization, and migration across hardware. This is a remote, part-time consulting engagement appealing to experts who enjoy hardware-close optimization.