GPU Kernels AI Task Auditor - Freelance AI Trainer Project

This position is listed on behalf of a partner company, who manages all applications and next steps. Our partner is looking for a GPU Kernels AI Task Auditor – Freelance AI Trainer Project based remotely in Brazil.

Join a specialized AI training project focused on improving the technical accuracy and reliability of advanced AI systems. As a GPU Kernels AI Task Auditor, you will evaluate technical tasks designed to train and assess AI models, ensuring they reflect realistic engineering challenges and meet rigorous standards. You will apply your expertise in GPU kernel development, parallel computing, and hardware-level optimization to identify technical flaws and performance issues. Your assessments will help strengthen the quality of AI training workflows and evaluation benchmarks. This freelance opportunity offers the flexibility of remote work and the chance to contribute to the development of more capable AI systems. It is particularly suited to experienced technical professionals who enjoy complex problem-solving and precise, evidence-based evaluation.

Accountabilities:
In this role, you will review GPU kernel tasks and provide actionable technical feedback to improve their correctness, practicality, and evaluation quality.

-

Evaluate GPU kernel tasks for technical accuracy, realism, solvability, and reproducibility.

-

Verify that tasks are supported by reliable tests, sound evaluation criteria, and clearly defined expected outcomes.

-

Identify memory allocation bugs, execution bottlenecks, and errors in parallel computing logic.

-

Analyze complex technical scenarios, troubleshoot issues, and assess whether proposed solutions are technically valid.

-

Provide clear, detailed, and actionable feedback to improve task quality and the reliability of AI training and evaluation workflows.

-

Apply your specialized expertise to the relevant technical domain, ensuring that assessments meet demanding engineering standards.

Requirements:

The ideal candidate brings substantial hands-on experience in GPU kernel development and the analytical ability to evaluate complex computing tasks rigorously. Candidates will be considered for a specific area of expertise based on their background; expertise across every listed technology is not required.

-

Demonstrable professional experience in GPU kernel development or a closely related technical specialty.

-

Deep technical knowledge of at least one relevant area, such as CUDA, Triton programming, parallel computing, or hardware-level performance optimization.

-

Strong analytical, critical-thinking, and problem-solving skills.

-

Ability to test, troubleshoot, and evaluate complex technical scenarios systematically.

-

Ability to identify implementation errors, performance limitations, and logical inconsistencies.

-

Strong written communication skills to explain technical findings and provide precise, actionable feedback.

-

Access to a secure computer and a reliable high-speed internet connection suitable for remote technical work.

Benefits:

-

Hourly compensation: Up to $60 per hour, with the exact rate determined by experience, expertise, and geographic location.

-

Remote flexibility: Work remotely.

-

Freelance arrangement: Contract-based work offering flexibility in how you organize your professional activities.

-

Technical impact: Contribute your GPU computing expertise to the improvement of AI training and evaluation systems.

-

Specialized work: Focus on the technical domain that best matches your experience.

-

Independent work setup: Use your own secure computer and high-speed internet connection.

-

Benefits eligibility: This is a freelance contractor position; company-sponsored health insurance, paid time off, and other employee benefits are not provided.