Build and optimize GPU software for LLM training tasks using CUDA, WebGPU, GLSL, and C++. This expert contractor role offers worldwide, part-time work at $60-$95 per hour.
Coding & Software
$60–$95/hr
Compensation
Worldwide
Eligibility
Expert
Experience
Jul 29, 2026
Posted
Open worldwide
OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. OpenTrain AI is hiring and contracting for this role, giving experts a place to apply their skills to cutting-edge AI work and build a lasting professional profile.
Creating an OpenTrain account is free, and qualified candidates can apply in minutes.
AI training is the human side of building artificial intelligence. People contribute technical expertise, examples, evaluations, and software that help modern AI systems learn and perform more effectively.
This project focuses on the engineering layer behind large language model training. Your GPU programming work can help improve the performance and efficiency of systems used to develop advanced AI.
OpenTrain is seeking an expert GPU Programming Software Engineer to design, implement, and optimize GPU-based software for LLM training tasks. The work centers on performance and architecture, including CUDA, WebGPU, or GLSL solutions, kernel and shader tuning, and C++ host-side integrations.
This is output-based contractor work for an engineer who can move quickly into GPU-related task delivery and produce work that meets defined project specifications.
You will contribute to defined GPU programming tasks supporting LLM training. The work requires strong attention to performance, architecture, and implementation quality.
This role is intended for an expert GPU programmer with advanced experience developing software that runs on NVIDIA chips. You should be able to move efficiently from onboarding into task work and deliver reliable implementations against defined requirements.
A portfolio or project history can help demonstrate the depth and quality of your GPU programming experience. Examples should show practical work with GPU software, performance optimization, or related technical systems.
AI training and data-labeling work is one of the fastest-growing ways to work in technology. OpenTrain helps contributors discover opportunities, showcase credible experience, and build a profile around the work they do across the industry.
For GPU specialists, this role offers a way to apply deep engineering expertise to the systems powering modern AI while developing a durable AI training portfolio.
Create a free OpenTrain account and apply for this GPU Programming Software Engineer contractor opportunity. Include portfolio or project examples that demonstrate your CUDA, WebGPU, GLSL, C++, kernel, shader, or GPU optimization experience.
Keep exploring
Use advanced CUDA, C++, GLSL, and WebGPU skills to optimize GPU kernels and evaluate performance across modern architectures. This remote contractor role offers $60-$100 per hour and 20+ hours per week.
Posted Aug 11, 2026
Evaluate GPU and accelerator kernel tasks for correctness, performance, compilation, and runtime validity. This fully remote US contract role pays $70 to $90 per hour and requires expertise across CUDA, Triton, NKI, or Pallas.
Posted Aug 27, 2026
Create reproducible reinforcement learning environments that test AI models on realistic software engineering workflows. Work globally for about 15 hours weekly at $100 to $150 per hour.
Posted Aug 4, 2026
Browse related job pages
Languages
