AI Red Team Engineer (C1 English)

OpenTrain is hiring multiple AI Red Team Engineers for remote, part-time contract work. You will design and run adversarial tests that probe how large language models, AI agents, and retrieval-augmented generation pipelines behave under attack.

You will follow detailed project guidelines and uphold strict ethical and safety standards while producing clear, reproducible security findings.

• Create and automate attack prompts and build repeatable test suites.
• Test prompt injection, jailbreaks, data exfiltration, function calling, and tool use.
• Evaluate LLM attack surfaces and apply the OWASP Top 10 for LLMs.
• Define scoring rubrics and grade model behavior.
• Document reproducible findings with risk ratings and propose mitigations.
• Contribute small scripts and utilities to scale testing.
• Use red-team frameworks such as garak or PyRIT.

This is a remote contractor role open worldwide. The work is part time and requires advanced English communication.

• Pay: $40 USD per hour.
• Hours: Less than 20 hours per week.
• Work arrangement: Remote, part-time contract.
• Language: Advanced English proficiency at C1 level.
• Experience: Intermediate level, with a strong cybersecurity and penetration testing background.
• Technical skills: Python, Bash, PowerShell, containerization, Docker, and CI/CD security.
• Testing experience: Web, API, network, and infrastructure security testing.
• LLM security knowledge: Prompt injection, jailbreaks, data exfiltration, LLM attack surfaces, and the OWASP Top 10 for LLMs.
• Assessment: Availability to complete a HackerRank assessment and a platform assessment as soon as possible.

Apply on OpenTrain. The employer reviews applications there.

AI training is the human work behind modern artificial intelligence, including testing model responses and finding unsafe or unreliable behavior. People with cybersecurity and other specialist skills help make AI systems safer and more useful.