AI Security Engineer Offensive and Defensive

You will review and penetration-test AI agents, design defenses against emerging AI threats, and build automated security detection tools. You will audit agent behavior, detect runtime anomalies, assess cloud-native attack surfaces, and implement defense-in-depth solutions.

Responsibilities

• Review and penetration-test AI agents
• Design and implement defenses against AI tool misuse, context poisoning, data and prompt poisoning, and prompt injection
• Develop AI-based automated security detection tools
• Enable automated alert triage and attack attribution
• Build agent behavior auditing and anomaly detection systems
• Identify abnormal agent runtime behavior chains
• Design and implement AI-driven security operations defenses and countermeasures
• Assess cloud-native agent runtime attack surfaces
• Design defense-in-depth solutions for Kubernetes, Docker, and cloud services

Requirements

• Have at least 3 years of security offense and defense experience, including penetration testing or red team practice
• Have at least 1 year of LLM security offense and defense experience
• Understand network architecture and cloud-native technology stacks, including Kubernetes, Docker container security, IAM, network isolation, and cloud-native threat models
• Build and modify AI agents using frameworks such as LangChain
• Demonstrate practical offensive capabilities against agents, including tool misuse, prompt injection, and context poisoning
• Know OWASP Top 10 for LLM and MITRE ATT&CK
• Be proficient in Python, Go, or C++
• Develop security offense and defense tools or extend open-source security platforms