Remote Freelance GPU Kernel Auditor for AI Training Tasks

Jobgether; Netherlands; Remote; EUR 44,000 - 74,000; Full time

Benefits offered by this job

Hourly rate up to $60

Remote work

Freelance contract

Technical impact

Independent work setup

No employer benefits

Job summary

Jobgether partners with a client to recruit a GPU Kernels AI Task Auditor for a freelance AI trainer project, remotely from the Netherlands. The role focuses on evaluating GPU kernel tasks for accuracy, realism, solvability, and reproducibility to strengthen AI training workflows and benchmarks.

You will verify tests and evaluation criteria, identify memory and parallel computing issues, and provide actionable feedback to improve task quality.

Qualifications

• Experience in GPU kernel development or closely related field.
• Deep knowledge of CUDA, Triton programming, parallel computing, or hardware-level optimization.
• Strong analytical and problem-solving abilities.
• Ability to systematically test, troubleshoot, and evaluate complex technical scenarios.
• Strong written communication to explain technical findings clearly.

Responsibilities

• Evaluate GPU kernel tasks for technical accuracy, realism, solvability, and reproducibility.
• Verify tasks are supported by reliable tests, sound evaluation criteria and defined outcomes.
• Identify memory allocation bugs, bottlenecks and parallel computing errors.
• Analyze complex scenarios, troubleshoot, and assess proposed solutions for validity.
• Provide clear, detailed feedback to improve task quality and AI training/evaluation workflows.
• Apply expertise to ensure assessments meet demanding engineering standards.

Skills

GPU kernel development

CUDA

Triton programming

Parallel computing

Hardware-level optimization

Analytical thinking

Problem solving

Written communication

Job description

Jobgether partners with a client to recruit a GPU Kernels AI Task Auditor for a freelance AI trainer project, remotely from the Netherlands. The role focuses on evaluating GPU kernel tasks for accuracy, realism, solvability, and reproducibility to strengthen AI training workflows and benchmarks.

You will verify tests and evaluation criteria, identify memory and parallel computing issues, and provide actionable feedback to improve task quality.