AI Jailbreak & Prompt-Injection Security Expert (AI Security)

Location: Remote, US

Company

hirify.global is engaging experts through micro1 for a customer initiative focused on AI safety and robustness.

What you will do

• Design methodologies for evaluating AI system safety across ethical jailbreaks, LLM red teaming, prompt injection, and tool-use abuse.
• Create cross-domain elicitation strategies to identify complex, multi-turn adversarial bypass patterns.
• Develop and maintain regression test suites for jailbreak susceptibility and prompt-injection vulnerabilities.
• Build evaluation frameworks to stress-test AI models against real-world adversarial threats.
• Collaborate with technical stakeholders to turn security findings into model safety and risk-mitigation improvements.
• Document methodologies, findings, and best practices in reports and presentations for technical and non-technical audiences.

Requirements

• 2+ years of expertise in adversarial machine learning, LLM red teaming, AI safety evaluation, or a closely related security domain.
• Proven experience researching, testing, or uncovering vulnerabilities involving ethical jailbreaks, prompt injection, tool-use abuse, or adversarial AI attacks.
• Advanced degree such as a PhD or MS in computer science, cybersecurity, machine learning, or a relevant discipline, or equivalent operational/professional background.
• High credibility or recognition in the AI security or adversarial ML community through published research, open-source tools, or conference presentations.
• Exceptional written and verbal communication skills, including clear documentation and collaborative problem-solving.

Nice to have

• Participation in multidisciplinary projects or cross-functional AI safety initiatives.
• Familiarity with current LLM architectures, prompt engineering techniques, and security assessment tools.

Culture & Benefits

• Contractor engagement for a customer initiative focused on AI safety and robustness.
• Remote work arrangement.
• No prior AI experience is required when relevant domain expertise is available.