Python Infrastructure Engineer, LLM Agent Tooling

You will build and maintain the infrastructure behind LLM training and agent evaluation. The work focuses on reusable repositories, secure sandboxes, task frameworks, scoring pipelines, and developer environments that help AI researchers iterate on agent tasks.

You will also help researchers use these tools and contribute to a collaborative engineering workflow.

• Build secure sandboxes for agent tasks.
• Create task frameworks and automated scoring pipelines.
• Develop reusable Python repositories and developer environments.
• Design CI/CD pipelines, with GitHub Actions preferred.
• Build or support FastAPI and Flask back ends.
• Use pytest-driven testing to check code and evaluation workflows.
• Use devcontainers, Makefiles, and other modern development tools.
• Guide AI researchers through the infrastructure and tooling.

This is a part-time contract role for candidates in the United States or Canada. The listed experience level is intermediate, and compensation is tiered by experience.

• Pay: $34 per hour for junior contributors, $37 per hour for middle-level contributors, and $42 per hour for senior contributors, in USD.
• Time: 20 or more hours per week.
• Location: Fully remote for candidates in the United States or Canada.
• Language: Clear English communication is required.
• Experience: At least five years of hands-on Python experience.
• Technical skills: Deep Linux and Docker knowledge, plus experience designing CI/CD pipelines.
• Back-end skills: Familiarity with FastAPI or Flask is required.
• Testing: Experience with pytest-driven testing is required.
• Tooling: Familiarity with devcontainers and Makefiles is required.
• Working style: A collaborative mindset and the ability to guide AI researchers are required.
• Bonus: Prior work on OpenTrain coding tasks is a strong plus.
• Experience level: Intermediate.
• Selection requirement: Short-listed candidates complete a timed HackerRank assessment and a platform coding test before recruiter interviews.

Apply on OpenTrain. The employer reviews applications there.

AI training work is the human side of building artificial intelligence. People create code, examples, and evaluations that help AI systems learn how to perform tasks, and experienced engineers are paid to make the tools and tests behind that work reliable. OpenTrain helps people find and build careers in this field.