Remote Freelance GPU Kernel Auditor for AI Tasks

Jobgether; Saudi Arabia; On-site; SAR 362,000 - 517,000; Full time

Benefits offered by this job

Competitive hourly rate

Fully remote

Project-based contract

Job summary

Jobgether is seeking a GPU Kernels AI Task Auditor - Freelance AI Trainer Project based in Saudi Arabia, offering fully remote work. You will audit GPU kernel tasks used to train and evaluate AI models, ensuring accuracy, realism, solvability, reproducibility, and robust testing criteria.

The role requires deep knowledge of GPU computing and memory management. The position emphasizes independent work with fast feedback on complex CUDA/Triton implementations, memory issues, and parallel

Qualifications

• Proven experience in GPU kernel development and optimization.
• Expertise in CUDA or Triton programming and parallel computing.
• Strong analytical and problem-solving skills for evaluating complex technical scenarios.
• Ability to work independently in a fully remote freelance environment.

Responsibilities

• Evaluate GPU kernel tasks for accuracy, realism, solvability, reproducibility, and robust tests.
• Rigorously test and troubleshoot GPU programming scenarios to validate behavior.
• Audit tasks for memory allocation, execution bottlenecks, synchronization issues, and parallel logic errors.
• Review CUDA, Triton, and other GPU kernel implementations for correctness and performance.
• Provide clear, actionable feedback and recommend improvements.
• Apply deep GPU computing knowledge to assess AI training/evaluation task quality.
• Work independently within a specialization while maintaining high technical accuracy.

Skills

GPU kernel development

CUDA

Triton

Parallel computing

Job description

Jobgether is seeking a GPU Kernels AI Task Auditor - Freelance AI Trainer Project based in Saudi Arabia, offering fully remote work. You will audit GPU kernel tasks used to train and evaluate AI models, ensuring accuracy, realism, solvability, reproducibility, and robust testing criteria.

The role requires deep knowledge of GPU computing and memory management. The position emphasizes independent work with fast feedback on complex CUDA/Triton implementations, memory issues, and parallel