AI Security Expert

Runpoint Partners; United States; Remote; USD 150,000 - 230,000; Full time

Benefits offered by this job

Remote role (US-friendly hours)

Contract-to-project opportunity

Job summary

Runpoint Partners seeks an AI Engineer to help make AI-enabled systems safer while maintaining productive workflows. This remote role emphasizes hands-on security testing, practical mitigations, and clear communication with engineering teams.

You will examine AI workflows, agent boundaries, and tool permissions, translating findings into actionable controls. Strong writing and speaking skills are essential for cross-team collaboration.

Qualifications

• Experience assessing or securing software, cloud environments, toolchains, or AI workflows.
• Knowledge of security testing techniques and modern tooling.
• Ability to reason about AI-specific risks and access controls.
• Hands-on approach with systems and tools, not only policy.
• Strong judgment on severity and actionable risk, with clear communication.

Responsibilities

• Deliberately test AI-enabled workflows to uncover realistic failure modes.
• Assess controls around permissions, sandboxing, secrets, and tool access.
• Identify vulnerabilities across code, dependencies, infra, and cloud access.
• Differentiate credible risks from demonstrations not translating to real attacks.
• Recommend mitigations, implement fixes, and verify problem resolution.
• Explain security trade-offs clearly to engineers and operators.

Skills

Security testing

Cloud security

AI risk assessment

Hands-on approach

Communication skills

Tools

CodeQL

Semgrep

Dependency scanning

Cloud IAM

Job description

AI Engineer Remote with US-friendly hours

Job details

Remote with US-friendly hours Eligibility

Senior Experience

Not specified Employment

About this role

Role overview

This role focuses on making AI-enabled systems safer without slowing useful work to a halt. You will investigate how AI workflows, agents, permissions, and integrations can fail, then turn those findings into practical controls that teams can actually adopt. The work combines hands-on security testing, technical judgment, and clear communication with people building and operating software.

Responsibilities

• Deliberately test agent workflows and AI-enabled systems to uncover realistic failure modes and exploit paths.
• Assess controls around permissions, sandboxing, secrets, approvals, and external tool access.
• Use security and code-quality tools to identify vulnerabilities across application code, dependencies, infrastructure, and cloud access.
• Distinguish credible risks from demonstrations that do not translate into a meaningful attack path.
• Recommend mitigations, help implement fixes, and verify that the changes resolve the underlying problem.
• Explain security trade-offs clearly, including when a solution can proceed with appropriate conditions rather than being rejected outright.

Requirements

• Practical experience assessing or securing software, cloud environments, developer tooling, or AI-enabled workflows.
• Working knowledge of techniques and tools such as CodeQL, Semgrep, dependency scanning, cloud IAM, or comparable security methods.
• Ability to reason about AI-specific risks, including agent boundaries, tool permissions, data exposure, and approval loops.
• A hands-on approach: comfortable using the systems and tools being evaluated, not only writing policy about them.
• Strong judgment about severity, exploitability, and the difference between theoretical concern and actionable risk.
• Clear written and verbal communication, with the ability to make recommendations useful to engineers and operators.

Benefits and work setup

• Remote role with working hours aligned to the United States.
• The engagement may begin with paid contractor work on a real project, with the possibility of expanding the relationship when the fit is strong.