AI Cybersecurity Engineer (Offensive Security & Threat Modeling)

About the role

As an AI Cybersecurity Engineer you will act as an internal adversary for a leading Egyptian financial institution. Working exclusively on a secure, high‑performance on‑premises environment, you will proactively threat‑model, simulate advanced AI attacks and manage algorithmic risks to protect the bank’s AI/ML pipelines and data assets.

Key responsibilities

• Lead predictive AI threat‑modeling exercises across AI/ML pipelines, data repositories and model training environments.
• Conduct offensive AI red‑team operations, including data poisoning, model inversion, membership inference, evasion attacks and prompt injection.
• Implement and operationalise the MITRE ATLAS framework, mapping vulnerabilities to ATLAS tactics and techniques.
• Assess and harden on‑premises infrastructure such as bare‑metal GPU clusters, Kubernetes/OpenShift container platforms and private data registries.
• Collaborate with security and data teams to design pre‑emptive guardrails before AI models move to production.

Required profile

• Strong offensive security mindset with proven experience in AI/ML threat modeling and red‑team simulations.
• Deep understanding of adversarial techniques targeting machine‑learning models.
• Familiarity with MITRE ATLAS and traditional MITRE ATT&CK frameworks.
• Hands‑on experience securing on‑premises GPU clusters and containerised environments.
• Ability to work independently within a high‑security, on‑site setting.

Required skills

• Python (for scripting attacks and analysis)
• Kubernetes
• OpenShift
• GPU clusters (bare‑metal)
• Private data registries
• MITRE ATLAS
• MITRE ATT&CK
• AI/ML pipeline security
• Adversarial attack techniques (data poisoning, model inversion, membership inference, evasion, prompt injection)