Cybersecurity Defence Analyst

The work

Help ensure AI security guidance is accurate, safe and practical. Evaluate defensive recommendations, challenge generated reasoning and identify mistakes that could cause harm. No prior AI or machine learning background is required.

• Review threat assessments, remediation advice, security settings and defensive strategies for technical accuracy and practical use.
• Flag dangerous errors, obsolete practices, configuration problems and misleading recommendations.
• Assess whether incident response plans, detection rules and hardening guidance would function in production.
• Create realistic security scenarios and demanding prompts to expose model blind spots.
• Provide structured feedback on completeness, safety and correctness.
• Check defensive scripts, code examples and tool settings against sound practices.
• Complete assignments independently and asynchronously.

Experience and skills

• Hands on defense experience in network security, endpoint protection, SIEM or SOC operations, incident response, vulnerability management, cloud security or identity and access management.
• Knowledge of MITRE ATT&CK, kill chain approaches and defensive responses.
• Ability to identify flaws in recommendations that could create operational harm.
• Clear, concise writing for technical explanations.
• Methodical review and reliable independent working habits.
• AI, ML or data science experience is not required.
• CISSP, CISM, CEH, Security+, CySA+, GIAC or equivalent certifications are useful.
• SOC, CSIRT or cybersecurity consulting work is preferred.
• Familiarity with firewalls, IDS/IPS, EDR, SIEM tools and vulnerability scanners is helpful.
• Penetration testing, red teaming, threat hunting, security documentation or runbook writing are valuable additions.
• Knowledge of NIST, ISO 27001, CIS Controls, Cyber Essentials or NCSC guidance is useful.
• AWS, Azure, GCP or container security and experience interacting with AI tools are further advantages.

Working arrangements

This is an hourly contract for 10 to 40 hours each week, carried out fully remotely. Work independently and asynchronously on a flexible schedule. Further assignments or an extension may become available as new projects begin.

The work supports research into safer AI security guidance and collaboration with other domain experts.