Cybersecurity/Code Red-Team AI Trainer, $65-$75/hour

Project Overview:

Join a growing community of professionals advancing the next wave of AI. As an AI Trainer, you’ll play a hands-on role by analyzing and providing feedback on data to improve LLM performance, helping ensure that the next generation of AI technology is accurate and trustworthy.

We are seeking a skilled Cybersecurity / Code Red-Team Expert to work as a project consultant in our AI Labor Marketplace. This is not a full-time employment position — you will be engaged as an expert project consultant on a contract basis.

Location: U.S.-based experts only

Engagement: Part-time, project-based expert evaluation work

Work Type: Remote

Project Summary:

Contributors will participate in authorized adversarial safety testing of vision-language models (VLMs), focusing on code and cybersecurity scenarios. The work involves creating technically credible image-and-text challenges designed to identify model safety failures, testing those challenges in a controlled environment, and documenting actual model responses with supporting evidence.

Contributors may also create legitimate-use contrast cases to identify unnecessary model refusals. All testing is authorized and controlled and uses safe, nonfunctional artifacts rather than live targets. The work may involve sensitive cybersecurity subject matter.

Consultant Engagement Terms:

This is a project-based consultant role. Consultants will be paid on a per-project basis; hourly rates are estimates based on anticipated completion time. Consultants control their own schedule, provide their own tools, and may simultaneously provide services to other vendors or employers (subject to those vendors’ allowances).

Responsibilities:

Contributors will:

- Design adversarial image-and-text challenges involving code and cybersecurity scenarios.
- Create safe technical visuals and nonfunctional artifacts for controlled VLM testing.
- Test model safeguards and document actual model responses.
- Identify and clearly substantiate potential safety failures with supporting evidence.
- Develop legitimate-use contrast cases to test for unnecessary model refusals.
- Read and reason about code and evaluate the cybersecurity implications of model outputs.
- Produce clear, technically accurate written documentation in English.
- Address technical or evidentiary issues identified during review.

Independent reviewer opportunities are also available for experts qualified to assess the technical accuracy, security implications, and supporting evidence of submitted findings.

Expected Outcomes:

- Technically credible adversarial VLM test cases.
- Appropriate image-and-text test materials using controlled, nonfunctional artifacts.
- Documented model responses and evidence supporting reported findings.
- Legitimate-use contrast cases where appropriate.
- Technically accurate submissions suitable for independent review.

Qualifications:

- Demonstrated cybersecurity, security engineering, or software engineering competence.
- Ability to read and reason about code and understand its security implications.
- Strong adversarial and analytical reasoning.
- Ability to document technical findings clearly in English.
- Relevant experience in one or more areas such as network security, web applications, binary analysis, memory safety, protocols, firmware, operating systems, malware, credentials, detection/evasion, or social engineering.
- Relevant backgrounds may include Red Teamer, Penetration Tester, Security Researcher, Application Security Engineer, Software/Systems Engineer, Reverse Engineer, Malware Analyst, or Incident Responder.
- Prior AI, image-based, or VLM red-teaming experience is a plus but is not required.
- No fixed degree or minimum-years-of-experience requirement; demonstrated practical competence is the primary consideration.

LinkedIn Data Security and Privacy Notice:

https://legal.linkedin.com/content/dam/legal/Global-Data-Privacy-Notice_LI_TalentCommunity.pdf