AI Safety Engineer: Guardrails, Adversarial Testing & Alignment

Hyphen Connect; Alice Springs; Presencial; AUD 120 000 - 180 000; Tempo integral

Resumo da oferta

Hyphen Connect is seeking an AI Safety Specialist to strengthen the security and robustness of language models deployed in real-world settings. You will conduct adversarial testing, implement guardrails, and contribute to safe tool integration while aligning behavior with ethical guidelines.

You will work on RLHF alignment pipelines, constitutional AI principles, and real-time filtering to prevent unsafe outputs in autonomous systems, collaborating with cross-functional teams in Australia.

Qualificações

• Background in cybersecurity, prompt engineering, or adversarial ML.
• Experience with jailbreak taxonomies and automated red-teaming frameworks.
• Strong analytical mindset for identifying edge cases.

Responsabilidades

• Conduct adversarial testing on LLMs and multimodal agents.
• Implement guardrails and real-time filtering for autonomous tool use.
• Develop constitutional AI principles and assist with RLHF alignment pipelines.

Conhecimentos

Cybersecurity

Prompt engineering

Adversarial ML

Jailbreak taxonomy

Automated red-teaming

Ferramentas

Red-teaming frameworks

Threat modeling tools

Descrição da oferta de emprego

Hyphen Connect is seeking an AI Safety Specialist to strengthen the security and robustness of language models deployed in real-world settings. You will conduct adversarial testing, implement guardrails, and contribute to safe tool integration while aligning behavior with ethical guidelines.

You will work on RLHF alignment pipelines, constitutional AI principles, and real-time filtering to prevent unsafe outputs in autonomous systems, collaborating with cross-functional teams in Australia.