GenAI CBRNE Security Red Team Specialist

Alice (Formerly ActiveFence); Imus; On-site; PHP 9,352,000 - 11,097,000; Full time

Benefits offered by this job

Health insurance

Job summary

Alice is seeking an experienced GenAI Red Team Expert to adversarially test safety guardrails for GenAI systems against high‑risk cyber threats in CBRNE domains. You will combine cybersecurity and domain knowledge to craft complex prompts, multi‑turn attacks, and model evaluations to prevent information leakage or capability misuse.

The role requires creative risk thinking, documentation of attack patterns, and ongoing collaboration with AI safety and model‑alignment teams.

Qualifications

• Advanced degree in a relevant technical field or equivalent experience.
• Deep knowledge of cyber threats, red teaming, and AI safety testing.
• Experience with adversarial prompt methods and model evaluation.
• Familiarity with critical infrastructure and OT environments.
• Strong documentation and reporting abilities.

Responsibilities

• Execute adversarial red‑teaming of GenAI models against cyber‑CBRNE scenarios.
• Design prompts, jailbreaks, and multi‑turn tests to probe safeguards.
• Create realistic cyber‑CBRNE attack scenarios and evaluate responses.
• Document failure modes, severity, and recommended mitigations.
• Develop repeatable test suites and evaluation rubrics.
• Stay current on emerging GenAI attack techniques and safety standards.
• Collaborate with safety teams to harden models and guardrails.

Skills

Cybersecurity

Red Team

GenAI Red Teaming

ICS/OT

Adversarial Mindset

Communication

Education

Advanced degree in Cybersecurity or related field

Job description

Alice is seeking an experienced GenAI Red Team Expert to adversarially test safety guardrails for GenAI systems against high‑risk cyber threats in CBRNE domains. You will combine cybersecurity and domain knowledge to craft complex prompts, multi‑turn attacks, and model evaluations to prevent information leakage or capability misuse.

The role requires creative risk thinking, documentation of attack patterns, and ongoing collaboration with AI safety and model‑alignment teams.