GPU Kernels AI Task Auditor - Freelance AI Trainer Project

Role overview

A freelance, remote contract position for technical specialists who audit GPU kernel tasks used to train and evaluate AI systems. The auditor's mission is to ensure these tasks are technically rigorous, realistic, and reproducible, providing expert feedback that raises the quality of AI training workflows in their specific area of expertise.

Responsibilities

- Review GPU kernel tasks for technical accuracy, realism, solvability, and reproducibility. - Assess accompanying tests and evaluation criteria for reliability and meaningfulness. - Identify memory allocation bugs, execution bottlenecks, and parallel computing logic errors within task designs. - Produce clear, actionable written feedback that task authors can act on. - Apply CUDA or Triton knowledge to judge whether hardware-level performance optimization claims are sound.

Requirements

- Demonstrable professional experience developing GPU kernels. - Working knowledge of CUDA and/or Triton programming. - Familiarity with parallel computing concepts and hardware-level performance optimization. - Strong analytical and problem-solving skills for testing and troubleshooting complex technical scenarios.

Benefits and work setup

Pay rate of $60 per hour, with the final rate determined after evaluating experience and geographic location. Engagement is a freelance contract, fully remote. Contractors supply their own secure computer and high-speed internet; company-sponsored health insurance and PTO do not apply.