AI Red Team Engineer — LLM Security & Pentesting

About OpenTrain

OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. We connect people to cutting-edge projects where they teach and shape how AI systems behave — from annotation and evaluation to adversarial testing — with flexible, remote work you can do around your life.

• Build experience in a fast-growing industry working directly on how state-of-the-art AI is trained and secured.
• Find part-time, remote contract work that values cybersecurity and specialist technical skills.

About AI Training and Red Teaming

AI training (data labeling, annotation, and human feedback) is how modern models learn. Red teaming and adversarial evaluation are a crucial subset: expert contributors design attacks and tests that reveal vulnerabilities so models can be hardened before deployment.

This project focuses on LLM and pipeline security — prompt injection, jailbreaks, data exfiltration, function-calling abuse, and RAG/agent weaknesses — and expects rigorous, ethical testing and clear, actionable reporting.

• Work influences real-world model safety and defenses.
• Most roles are remote and flexible, making this an excellent part-time technical opportunity.

The Role

We’re hiring multiple AI Red Team Engineers to design and execute adversarial evaluations of LLMs, agents, and RAG pipelines in a remote, part-time contract capacity. You will follow detailed project guidance, uphold strict ethical and safety standards, and deliver reproducible findings and remediation suggestions.

This is a contractor role, less than 20 hours per week, paid at USD 40 per hour. You must be available to complete a HackerRank plus platform assessment immediately after screening and communicate clearly in advanced (C1) English.

• Employment type: Contractor, Part-time.
• Time requirement: Less than 20 hours/week.
• Compensation: $40/hour (USD).
• Language: Advanced English (C1) required.
• Assessment: HackerRank + platform test required ASAP after screening.

What You’ll Do

You will plan, automate, and run adversarial test suites that probe models, tool/function calling, and RAG components; grade model responses; and document results with clear risk ratings and mitigations.

Work will include both manual red-teaming and building small utilities to scale tests, plus writing concise, reproducible reports and scoring rubrics for graders or automatic evaluation.

• Craft and automate adversarial prompts and attack suites for LLMs, agents, and RAG pipelines.
• Probe function-calling and tool use to identify abuse or data leakage.
• Define scoring rubrics and grade model behaviors consistently.
• Document reproducible findings with risk ratings and recommended mitigations.
• Contribute small scripts/utilities to scale automated testing.
• Collaborate with remote teams and follow strict ethical testing guidelines.

Requirements

You must meet the below mandatory requirements. Applicants who cannot demonstrate these skills and immediate test availability should not apply.

Restricted locations: candidates from certain countries, states and territories are ineligible — see the full list below and confirm your eligibility when you apply.

• Bachelor’s or Master’s in Computer Science, Software Engineering, Cybersecurity, Digital Forensics, or a related field.
• Hands-on penetration testing experience across web, API, network, and infrastructure; familiarity with cloud and container security.
• Strong scripting/automation skills in Python, Bash, or PowerShell.
• Experience with containerization and CI/CD security tools (e.g., Docker) and secure SDLC practices.
• Deep knowledge of LLM vulnerabilities (prompt injection, jailbreaks, data leakage) and the OWASP Top 10 for LLMs.
• Familiarity with AI red-teaming/eval frameworks (e.g., garak, PyRIT) and experience evaluating LLMs, agents, and RAG pipelines.
• Offensive exploitation and reverse engineering experience (e.g., Ghidra or equivalent); OS security skills such as Linux privilege escalation and Windows internals.
• Ability to write clear rubrics, adversarial prompts, and concise reports; familiarity with secure coding practices for full-stack review.
• Availability to complete HackerRank + platform assessment immediately after screening.
• Advanced English (C1) for clear written and spoken communication.

Nice-to-Have

These are not required but strengthen your application and may help you move faster through selection.

• Prior experience at leading AI or security organizations or shipped LLM security work.
• Experience building evaluation tooling or automations for large-scale model testing.
• Familiarity with secure CI/CD pipelines and infrastructure-as-code security.

Who Should Apply

Apply if you are a hands-on offensive security engineer with practical pentesting experience, fluency in scripting and container/cloud tooling, and real-world knowledge of LLM attack surfaces. Ideal candidates can translate technical findings into concise risk-rated reports and move quickly through technical assessments.

• You enjoy adversarial thinking, automation, and clear technical communication.
• You can work independently in a remote, part-time contractor setup and meet quick assessment deadlines.

How It Works & How to Apply

Apply via OpenTrain and complete the platform screening. Qualified candidates will be asked to finish a HackerRank test plus a platform assessment immediately after screening. Successful applicants will onboard as contractors and work remotely under project-specific guidelines and ethical constraints.

When you apply, be prepared to show examples of past pentesting or red-team work, code snippets or scripts you’ve written for testing, and confirm you meet the educational and technical requirements.

• Submit your profile and application on OpenTrain.
• Complete HackerRank + platform assessment as requested — must be available ASAP.
• Contract starts after selection and onboarding; work remotely and part-time (<20 hrs/week).

Location Eligibility (Ineligible List)

Candidates from the following countries, territories, or regions are ineligible for this project. If your country or state is listed here, do not apply.

Ineligible: Iran, Cuba, North Korea, Syria, Sudan, Venezuela, Myanmar; Switzerland; China, Taiwan; Kenya; Armenia, Israel, Kazakhstan, UAE, Netherlands, Serbia, Kyrgyzstan, Turkey, Uzbekistan, Belarus, Russia, Ukraine, Abkhazia, South Ossetia; United States (restricted states): Alaska, Arkansas, California, Connecticut, Delaware, Georgia, Hawaii, Illinois, Indiana, Kansas, Louisiana, Maine, Maryland, Massachusetts, Nebraska, Nevada, New Hampshire, New Jersey, New Mexico, Ohio, Oregon, Tennessee, Utah, Vermont, Washington, West Virginia; and these territories/regions: Antarctica, Aruba, Åland Islands, Saint Barthélemy, Bonaire, Sint Eustatius and Saba, Bouvet Island, Cocos (Keeling) Islands, Democratic Republic of the Congo, Cook Islands, Christmas Island, Western Sahara, Falkland Islands (Malvinas), French Guiana, Guadeloupe, South Georgia and the South Sandwich Islands, Heard Island and McDonald Islands, British Indian Ocean Territory, Northern Mariana Islands, Martinique, New Caledonia, Norfolk Island, Niue, French Polynesia, Saint Pierre and Miquelon, Pitcairn, Réunion, Saint Helena, Ascension and Tristan da Cunha, Svalbard and Jan Mayen, Sint Maarten (Dutch part), French Southern Territories, Tokelau, United States Minor Outlying Islands, Holy See, Virgin Islands (British), Wallis and Futuna, Mayotte.