Remote AI Security Analyst: Frontier AI Safety Testing

Alignerr Corp.; Greater London; Remote; GBP 83,000 - 165,000; Part time

Job summary

Alignerr is seeking an AI Security Analyst to probe frontier AI models, identifying vulnerabilities, and recommending mitigations. This is a remote, hourly contract role with flexible scheduling, designed for security professionals curious about how AI systems can be exploited and protected.

You will apply threat modeling frameworks to emerging AI technologies and deployment contexts, review prompt injection scenarios, data leakage, and model abuse, and work independently on task-based

Qualifications

• Background in cybersecurity, information security, or similar field.
• Strong understanding of security threat modeling for AI/LLM systems.
• Curious and analytical mindset; ability to assess complex systems.
• Precise and methodical when evaluating risks and implications.
• Self-motivated and reliable for task-based assignments.

Responsibilities

• Analyze real-world AI and LLM security scenarios to understand behavior under adversarial conditions.
• Review cases involving prompt injection, data leakage, model abuse, and system misuse.
• Classify vulnerabilities and recommend mitigations based on impact and likelihood.
• Apply structured threat modeling frameworks to emerging AI technologies and deployment contexts.
• Help evaluate and improve AI system behavior to remain safe and resistant to exploitation.
• Work independently and asynchronously on tasks.

Skills

Cybersecurity background

Threat modeling

Analytical mindset

Precise methodology

Independent work

Tools

MITRE ATT&CK

OWASP

NIST

Job description

Alignerr is seeking an AI Security Analyst to probe frontier AI models, identifying vulnerabilities, and recommending mitigations. This is a remote, hourly contract role with flexible scheduling, designed for security professionals curious about how AI systems can be exploited and protected.

You will apply threat modeling frameworks to emerging AI technologies and deployment contexts, review prompt injection scenarios, data leakage, and model abuse, and work independently on task-based