Python Backend Engineer for AI Agent Data

You will build Python backend applications that reproduce the workflows of tools such as Slack, Linear, Jira, Notion, Gmail, and wikis. You will also help create realistic tasks and evaluation methods that show whether AI agents complete backend work correctly.

The role combines software development with careful review. You will use AI coding assistants during development, testing, and validation, then check that connectors and evaluation tasks work as intended.

• Build and extend connectors for cloud software tools.
• Test and validate connectors so they behave faithfully.
• Deliver new connectors within agreed timeframes.
• Study data and workflows to find opportunities for multi-step task development.
• Write realistic tasks for evaluating AI agents.
• Check task quality and correctness through structured quality assurance.
• Create rubrics that distinguish correct, partly correct, and deficient work.
• Take part in onboarding, calibration, and quality reviews.

What it pays and takes

This is a remote, four-week contractor engagement. The work is open worldwide and requires English fluency.

• $300 per approved task.
• At least one completed and approved task per day.
• 20 or more hours per week.
• Additional tasks may be available based on project capacity.
• At least three years of backend development experience.
• Strong Python skills, including FastAPI, Flask, or Django.
• Experience building scalable backends, REST APIs, and microservices.
• Practical knowledge of Git, Docker, pipelines, testing, and code-quality practices.
• Daily experience with AI coding assistants such as Claude Code, Codex, Cursor, GitHub Copilot, or similar tools.
• Strong attention to detail, sound technical judgment, and the ability to follow project guidelines.

How it works

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

About AI training work

AI training is the human work behind modern AI systems, including creating tasks, reviewing outputs, and rating whether software or model responses are correct. People with strong technical experience are needed because their judgment helps make these evaluations accurate and useful.