Remote AI Safety Red Team Engineer

Obsidian; Johannesburg; Presencial; ZAR 1.457.000 - 2.104.000; Jornada completa

Descripción de la vacante

Mercor is assembling a remote red team to probe AI models with adversarial inputs, surface vulnerabilities, and generate data to improve safety. This work is text-based and involves reviewing outputs touching sensitive topics with clear guidelines and wellness resources.

You’ll annotate failures, follow taxonomies, and document reproducible reports and attack cases for customers to act on, collaborating across time zones and teams.

Formación

• Red team background with AI adversarial work or cybersecurity experience.
• Ability to apply frameworks or benchmarks to testing.
• Clear, concise communication of risks to technical and non-technical stakeholders.

Responsabilidades

• Red team conversational AI models and agents: jailbreaks, prompt injections, misuse cases, bias exploitation, multi-turn manipulation.
• Generate high-quality human data: annotate failures, classify vulnerabilities, and flag systemic risks.
• Apply structure: follow taxonomies, benchmarks, and playbooks to keep testing consistent.
• Document reproducibly: produce reports, datasets, and attack cases customers can act on.

Conocimientos

Red team experience

Adversarial AI

Structured testing frameworks

Communication skills

Adaptable to multiple projects

Descripción del empleo

Mercor is assembling a remote red team to probe AI models with adversarial inputs, surface vulnerabilities, and generate data to improve safety. This work is text-based and involves reviewing outputs touching sensitive topics with clear guidelines and wellness resources.

You’ll annotate failures, follow taxonomies, and document reproducible reports and attack cases for customers to act on, collaborating across time zones and teams.