Remote AI Safety Red Team Specialist

Obsidian; Dublin; Remote; EUR 60,000 - 90,000; Full time

Job summary

Mercor is assembling a remote red team to probe AI models and surface vulnerabilities. You’ll work on text-based adversarial tasks aimed at improving safety, with topics clearly communicated and optional participation in high-sensitivity projects.

You'll generate high-quality human data and follow established playbooks to ensure reproducibility. You bring prior red teaming experience, a curious, adversarial mindset, and strong communication skills.

Qualifications

• Prior red teaming experience in AI or cybersecurity.
• Ability to identify vulnerabilities and document risks clearly.
• Structured approach using frameworks or benchmarks; analytical mindset.
• Excellent written and verbal communication for technical and non-technical audiences.
• Adaptable to moving across projects and customers.

Responsibilities

• Red team conversational AI models and agents, including jailbreaks and prompt injections.
• Annotate failures, classify vulnerabilities, and flag systemic risks.
• Apply structure using taxonomies, benchmarks, and playbooks for consistency.
• Document reproducibly with reports, datasets, and attack cases for customers.

Skills

Red teaming

Adversarial ML

Cybersecurity

Socio-technical risk

Structured thinking

Clear communication

Tools

Benchmarks

Playbooks

Job description

Mercor is assembling a remote red team to probe AI models and surface vulnerabilities. You’ll work on text-based adversarial tasks aimed at improving safety, with topics clearly communicated and optional participation in high-sensitivity projects.

You'll generate high-quality human data and follow established playbooks to ensure reproducibility. You bring prior red teaming experience, a curious, adversarial mindset, and strong communication skills.