Remote AI Security Analyst — Defending Frontier Models

Alignerr; New York (NY); On-site; USD 83,000 - 165,000; Full time

Job summary

Alignerr is seeking an AI Security Analyst for a remote hourly contract to probe frontier AI models for vulnerabilities and improve safety and reliability at scale.

You will analyze real-world AI scenarios, review prompt injection and data leakage, classify issues by impact, and recommend mitigations. You’ll work independently, on a flexible schedule, with potential for ongoing projects and extensions as new AI security challenges emerge. Strong threat modeling background preferred.

Qualifications

• Background in cybersecurity, information security, or a closely related field.
• Strong understanding of modern security threat modeling and AI applications.
• Analytically sharp and precise when evaluating complex systems.
• Able to work through ambiguous scenarios without a defined playbook.
• Self-motivated and consistent when working independently.

Responsibilities

• Analyze real-world AI and LLM security scenarios to understand model behavior under adversarial conditions.
• Review and evaluate cases involving prompt injection, data leakage, model abuse, and system misuse.
• Classify security issues by impact and likelihood; recommend mitigations.
• Help evaluate and improve AI system behavior for safety, reliability, and security alignment.
• Work independently and asynchronously on your own schedule.

Skills

Cybersecurity

Threat modeling

Analytical thinking

Ambiguity tolerance

Independent work

Job description

Alignerr is seeking an AI Security Analyst for a remote hourly contract to probe frontier AI models for vulnerabilities and improve safety and reliability at scale.

You will analyze real-world AI scenarios, review prompt injection and data leakage, classify issues by impact, and recommend mitigations. You’ll work independently, on a flexible schedule, with potential for ongoing projects and extensions as new AI security challenges emerge. Strong threat modeling background preferred.