About the job
Please submit your CV in English and indicate your English proficiency level.
Mindrift connects skilled professionals with project-based AI work for top technology companies. The focus is on testing, evaluating, and improving AI systems. This is a contract position tied to specific projects, not a permanent full-time role.
Role overview
This Senior Python Systems Developer position centers on functional testing for large and complex codebases. The work involves designing and running black-box tests, setting up reproducible environments with Docker, and using large language models (LLMs) to enhance and automate development tasks. The ideal candidate is comfortable with Linux, Docker, and can interpret code in languages like C, Rust, or Go. Familiarity with iterative development tools such as Roo Code or Claude Code is valued, as is experience with migration tasks.
What you will do
- Design and implement functional black-box tests for large codebases in various languages.
- Set up and maintain Docker environments to support reproducible builds and test runs across platforms.
- Track code coverage and configure automated scoring to match industry standards.
- Leverage LLMs (including Roo Code and Claude) to automate repetitive steps and improve code quality.
Requirements
- Minimum 5 years as a Software Engineer, primarily working with Python.
- Strong hands-on experience with pytest (including fixtures, session-scoped tests, and timeouts) and black-box testing for CLI tools.
- Expert-level knowledge of Docker, including writing reproducible Dockerfiles, managing user contexts, and workspace security.
- Advanced Linux and Bash scripting skills, with the ability to debug inside containers.
- Familiarity with modern Python tooling such as uv, pyproject.toml, and packaging.
- Ability to read and interpret code in C, C++, Rust, or Go, with LLM assistance.
- Experience using LLMs (Claude Code, Roo Code, Cursor) for iterative development and generating tests.
- English proficiency at B2 level or higher.
Preferred
- Background with agent evaluation platforms and MCP CLI.
Tools and technologies
Python (pytest, uv, Pillow), Docker, Bash, Git Submodules, C/C++/Rust/Go (for code reading), Dagger, GitHub Codespaces, LLMs (Claude Code, Roo Code, Cursor), coverage.py, gcov, kcov.
Benefits and compensation
- Freelance, project-based work through the Mindrift platform (powered by Toloka AI).
- Fully remote and flexible: set your own schedule and contribution level (20-30 hours per week).
- Project-based pay. AI trainers on this project can earn up to $80 per hour.
Location: Remote , San Antonio, Texas, United States

