GPU Kernel and ML Systems Researcher (Coding Tasks)

You will create self-contained technical coding tasks for research-focused evaluation in ML systems and GPU kernels. You will also review AI model outputs and provide feedback grounded in your research expertise.

Your work may cover CUDA, Triton, CUTLASS kernels, inference serving, and ML compilers. You will work with a project lead to keep tasks accurate, rigorous, and suitable for technical research evaluation.

• Build technical research coding tasks involving ML systems and GPU kernels.
• Review AI model outputs against the standards of ML systems research.
• Provide technically grounded feedback on model performance and technical accuracy.
• Apply your research perspective to self-contained technical work.

What it pays and takes

This is remote, asynchronous contract work designed to fit around your research.

• Pay: $200 per approved task, typically taking about four hours.
• Work setup: Remote, asynchronous contract work.
• Location: Open to candidates in the listed countries.
• Language: Professional fluency in English.
• Research: Published work in ML systems and GPU kernels, including at venues such as MLSys, OSDI, SOSP, PPoPP, SC, or comparable venues.
• Education: A completed or in-progress PhD in computer science or a related field.
• Technical depth: Demonstrated expertise in CUDA, Triton, CUTLASS kernels, inference serving, or ML compilers.
• Research artifacts: Released research code through GitHub or another research artifact.
• Helpful background: First-author publications, Docker, reproducible research environments, and research engineering that combines published work with shipped code.

How it works

Apply on OpenTrain with your resume, then complete the application on the hiring site.

About AI training work

AI training work is the human side of building artificial intelligence. People prepare examples, review model responses, and evaluate technical outputs so AI systems can improve, and experienced specialists are needed when the work requires deep subject knowledge. OpenTrain helps people find these projects, build a profile, and apply in minutes.