Job Summary
We are seeking an AI Red Team Engineer to offensively test AI systems and the security controls surrounding their use. The role will operate from an attacker's perspective to pressure-test AI solutions in non-production environments, identify AI-specific weaknesses, and translate findings into practical guidance that strengthens guardrails, architecture, standards, and development practices. The role will focus specifically on offensive testing of AI applications and systems and will serve as a constructive partner to application and AI development teams.

Key Responsibilities
• Conduct offensive security testing of AI applications, model interactions, and AI-enabled systems in non-production environments.
• Perform prompt-level attacks, adversarial input testing, guardrail bypass testing, and assessment of model-specific weaknesses.
• Research and apply emerging AI attack techniques and evaluate their potential impact on enterprise AI environments.
• Assess whether newly discovered industry exploits and attack techniques are applicable to AI systems and architectures.
• Independently select appropriate testing methods, tools, and techniques based on the architecture and behavior of the AI system.
• Analyze AI systems from an attacker's perspective to identify meaningful exploitation paths and weaknesses.
• Evaluate LLM behavior, AI architectures, platform differences, prompt injection risks, model exploitation opportunities, and AI guardrails.
• Document findings and communicate security weaknesses clearly to technical and non-technical stakeholders.
• Participate in remediation discussions with application and AI development teams.
• Recommend controls, stronger prompts, input restrictions, architecture patterns, standards, and security best practices.
• Provide actionable guidance that enables development teams to strengthen AI security while retaining implementation ownership.
• Maintain awareness of the rapidly evolving AI threat landscape and emerging attack methodologies.
• Use appropriate AI-enabled security tools and techniques to enhance offensive testing capabilities.
• Collaborate with broader red team, infrastructure security, application security, AI engineering, and development teams as appropriate.

Required Qualifications
• Strong foundation in offensive security or red team operations combined with deep, practical AI knowledge.
• Demonstrated hands-on experience attacking AI systems, rather than solely developing AI solutions or using AI-assisted security tools.
• Strong understanding of Large Language Models (LLMs), model behavior, AI architectures, platform differences, and AI security risks.
• Hands-on experience with adversarial testing, prompt injection, model exploitation, guardrail bypasses, or related AI attack techniques.
• Ability to independently select testing methodologies and tools based on the target AI system.
• Ability to infer system architecture and identify meaningful exploitation paths.
• Strong research skills and awareness of emerging AI security threats and attack techniques.
• Strong communication skills with the ability to clearly explain findings and provide actionable security recommendations.
• Experience working across both AI and offensive security disciplines.
• Relevant experience may include red team, application security, DevSecOps, security research, AI engineering, or adversarial machine learning, provided there is demonstrated depth across AI and offensive security.

Preferred Qualifications
• Experience with AI red teaming, adversarial AI, or offensive testing of LLM-based applications.
• Experience assessing AI guardrails, prompt security, model interactions, and AI-specific attack surfaces.
• Experience with application security, DevSecOps, security research, or adversarial machine learning.
• Certifications such as OSCP may support general offensive security expertise but are not considered a substitute for hands-on AI red team experience.
• Experience researching and evaluating newly emerging AI exploits and attack methodologies.