Expert Talent Pool, Safety Testing (Expert, 5+ Years)

Clera · remote

RemotoTempo integral

Publicada há 1 hora

Sobre a vaga

About the Role

Join a curated expert pool supporting AI safety testing projects. You will lead evaluations across security, fairness, and specialized safety domains, helping identify and address model risks before deployment.

What You'll Do

  • Design and lead red teaming campaigns to identify attack vectors in high-risk model capabilities.

  • Develop jailbreak techniques and document them for reproducibility and knowledge sharing.

  • Test agents for prompt injection vulnerabilities when processing untrusted inputs.

  • Probe models with tool-use capabilities, including code execution, browsing, and messaging, for misuse scenarios.

  • Design bias testing frameworks and assess fairness across demographic groups and languages.

  • Conduct safety audits against written policies and document compliance gaps.

  • Write content moderation guidelines, review edge cases, and assess other moderators' work.

  • Rate finding severity and recommend appropriate fixes.

What We're Looking For

  • At least five years of experience in security testing, trust and safety, policy, fairness research, adversarial machine learning, or a relevant domain such as cyber, biological, chemical, or child safety, with hands-on AI safety testing experience.

  • Experience designing test plans, threat models, red team plans, or audits that others carried out.

  • Experience reviewing or leading testers or moderators.

  • Hands-on experience testing AI systems for red teaming, jailbreaks, prompt injection, or tool misuse.

  • Ability to rate findings by severity and write recommended fixes.

  • Familiarity with tools such as Burp Suite, Garak, or PyRIT, or with the OWASP LLM Top 10.

  • Strong written English and ability to work with sensitive content under clear ethical guidelines.

Compensation & Benefits

$75 to $175 per hour, depending on project, field, and experience. Payment is based on hours worked.

Location

Fully remote. Projects may range from one week to several months.

Leu e ficou na dúvida? A nota responde.

Ver minha nota