Python AI Safety Testing Engineer

Mindrift; United Kingdom; On-site; GBP 46,610 - 77,684; Full time

Job summary

Mindrift connects specialists with project-based AI opportunities for leading tech companies, focusing on testing, evaluating, and improving AI systems. Frontier coding agents design tasks where the safe path is required, and tests catch unsafe shortcuts, ensuring the evaluation is fair and robust.

Participation is project-based, not permanent employment. Tasks involve building realistic environments, designing tests, and iterating based on QA feedback to maintain code quality and safety

Qualifications

• 4–5+ years in software development.
• Proficient with Python and JavaScript/TypeScript.
• Strong test design for functional and integration tests.
• Experience with coding agents (Claude Code, Codex, Copilot CLI).
• Familiar with GitHub PRs and CI workflows.
• English proficiency at least B2+.

Responsibilities

• Design tasks that reveal safe vs unsafe completions.
• Write tests to catch shortcuts and ensure compliance with scope.
• Evaluate agent solutions and provide QA feedback.
• Iterate on tasks and tests for fairness and robustness.

Skills

4–5+ years experience

Python

JavaScript/TypeScript

Test design

Coding agents experience

GitHub PRs & CI

English B2+

Tools

Claude Code

GitHub Copilot CLI

Codex

Job description

Mindrift connects specialists with project-based AI opportunities for leading tech companies, focusing on testing, evaluating, and improving AI systems. Frontier coding agents design tasks where the safe path is required, and tests catch unsafe shortcuts, ensuring the evaluation is fair and robust.

Participation is project-based, not permanent employment. Tasks involve building realistic environments, designing tests, and iterating based on QA feedback to maintain code quality and safety