Red Team Lead, Offensive Cybersecurity for AI Evaluation

About OpenTrain

OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. We help people discover projects, build a unified portfolio, and grow durable freelance careers teaching AI. Creating an OpenTrain account is free.

About AI training and red teaming

AI training (data labeling and human evaluation) is the human side of building intelligent systems. Contributors create, evaluate, and refine the examples and tests that shape model behavior — including offensive and defensive security evaluations that help models resist misuse.

This role focuses on offensive cyber work used to evaluate and harden AI systems: designing safe proxy tasks, scoring attacker sophistication, and documenting findings so models and engineering teams can improve security posture.

The role

OpenTrain is recruiting a Red Team Lead to design and validate offensive-security evaluation work that supports AI training and model evaluation. You will develop cyber-capability taxonomies, create safe proxy attack tasks, produce scoring rubrics, review red-team benchmarks, and write clear technical documentation for both technical and non-technical audiences.

This is a remote contractor, part-time role with a minimum commitment of 20+ hours per week. Work is asynchronous and worldwide; English proficiency is required.

• Employment type: Contractor, part-time
• Time requirement: 20+ hours/week (asynchronous collaboration)
• Location: Remote, worldwide (English required)
• Pay: $50–$90 USD per hour

What you'll do

• Build taxonomies for cyber-capability tasks and attack stages to structure evaluation work.
• Design and validate offensive security evaluation frameworks that measure attack coverage and impact.
• Create safe proxy tasks that simulate advanced attack vectors while enforcing ethical boundaries.
• Write scoring rubrics for attack sophistication, coverage, and impact to support consistent ratings.
• Review and strengthen red-team benchmarks to address evolving security risks and exploit chains.
• Document methodologies and produce technical write-ups for both technical and non-technical stakeholders.
• Collaborate asynchronously with project stakeholders and incorporate feedback into deliverables.

Requirements

Candidates must preserve safety and ethics while working on adversarial security tasks. OpenTrain expects rigorous handling of sensitive subject matter and adherence to agreed boundaries and controls.

• 5+ years hands-on experience in offensive cybersecurity, red teaming, exploit development, or vulnerability research.
• Strong experience with exploit chains, malware analysis, cloud or application security, or social engineering.
• Proven ability to design evaluation frameworks and scoring rubrics for security work.
• Clear written and verbal communication for technical and non-technical audiences.
• Familiarity with ethical boundaries in adversarial security work; certifications or advanced degrees are valued.
• Experience building cybersecurity benchmarking or evaluation frameworks is a plus.

Technical details of the engagement

This project involves text-based evaluation tasks used for RLHF and structured evaluation ratings. You will produce frameworks, rubrics, and documentation rather than executing real-world attacks. All work must follow ethical safeguards and project-specific constraints.

• Data type: Text
• Labeling work: RLHF, evaluation rating
• Payment type: Hourly (USD $50–$90/hr)

Who should apply and how it works

Apply if you are an experienced offensive security professional who wants to shape how AI systems are tested and hardened. This is ideal for practitioners who enjoy writing clear methods, building benchmarks, and collaborating remotely on safety-focused work.

To apply, create an OpenTrain account, complete your profile, and submit your application for the Red Team Lead role. OpenTrain will manage contracting and project details; selected contributors will work asynchronously and deliver frameworks, tests, and documentation according to project milestones.

• Great for senior red teamers, exploit developers, and vulnerability researchers who want flexible, remote work.
• OpenTrain is the hiring organization and manages payments and contracts for this role.