Principal AI Red Team Operator

Jobtailor; Deutschland; Vor Ort; EUR 120.000 - 180.000; Vollzeit

Zusammenfassung

Jobtailor is seeking an experienced AI red team professional to build and operationalize adversarial testing across enterprise AI systems. You will drive manual and automated assessments of LLMs, RAGs, and agents, focusing on prompt injections, data poisoning, and model exploitation.

Present risk insights to leadership and guide remediation. The role requires deep AI safety/security expertise, strong Python skills, and cloud platform experience.

Qualifikationen

• Bachelor's degree or four or more years of relevant work experience.
• Six or more years of relevant work experience required, demonstrated through job-related work experience, military experience, or specialized training or education (non-collegiate).
• Experience conducting AI evaluations or developing AI Safety/Security systems.
• Experience using adversarial prompting strategies such as role play and DAN to compromise enterprise AI systems.
• Experience using non-prompt strategies such as data poisoning and embedding inversion to compromise AI systems.
• Strong understanding of AI model architectures and development.
• Strong understanding of cybersecurity principles and practices.
• Experience in traditional web application testing.
• Experience using local LLMs for prototyping and experimentation with GenAI, such as Ollama and LlamaFiles.
• Experience training and educating others on Generative AI tools.
• Knowledge of programming languages such as Python.
• Experience with cloud computing platforms such as GCP, AWS, or Azure.
• Demonstrated continual learning around Artificial Intelligence.
• Ability to work from home and at least three days per week in the office.
• Scheduled weekly hours: 40

Aufgaben

• Build, refine, and operationalize AI red team frameworks, playbooks, and automated testing tools aligned with industry standards such as MITRE ATLAS and OWASP
• Plan and conduct complex manual and automated adversarial assessments on enterprise AI systems, including LLMs, RAG architectures, and agents
• Target prompt injection, data poisoning, and model exploitation during AI red team assessments
• Present findings and risk insights to technical engineering teams and senior leadership
• Provide actionable remediation guidance to secure the AI lifecycle
• Track evolving GenAI vulnerabilities and experiment with novel attack vectors
• Continuously enhance testing capabilities to stay ahead of adversaries
• Lead development of testing frameworks, playbooks, toolsets, and metrics to scale Verizon's AI Red Teaming capability

Kenntnisse

AI Red Teaming

Adversarial Testing

Python Programming

Cybersecurity Principles

Cloud Platforms Experience

Web App Testing

LLMs / GenAI

Ausbildung

Bachelor's degree

Tools

MITRE ATLAS

OWASP

GCP

AWS

Azure

Ollama

LlamaFiles

Jobbeschreibung

• Build, refine, and operationalize AI red team frameworks, playbooks, and automated testing tools aligned with industry standards such as MITRE ATLAS and OWASP
• Plan and conduct complex manual and automated adversarial assessments on enterprise AI systems, including LLMs, RAG architectures, and agents
• Target prompt injection, data poisoning, and model exploitation during AI red team assessments
• Present findings and risk insights to technical engineering teams and senior leadership
• Provide actionable remediation guidance to secure the AI lifecycle
• Track evolving GenAI vulnerabilities and experiment with novel attack vectors
• Continuously enhance testing capabilities to stay ahead of adversaries
• Lead development of testing frameworks, playbooks, toolsets, and metrics to scale Verizon's AI Red Teaming capability

Requirements

• Bachelor's degree or four or more years of relevant work experience
• Six or more years of relevant work experience required, demonstrated through job-related work experience, military experience, or specialized training or education (non-collegiate)
• Experience conducting AI evaluations or developing AI Safety/Security systems
• Experience using adversarial prompting strategies such as role play and DAN to compromise enterprise AI systems
• Experience using non-prompt strategies such as data poisoning and embedding inversion to compromise AI systems
• Strong understanding of AI model architectures and development
• Strong understanding of cybersecurity principles and practices
• Experience in traditional web application testing
• Experience using local LLMs for prototyping and experimentation with GenAI, such as Ollama and LlamaFiles
• Experience training and educating others on Generative AI tools
• Knowledge of programming languages such as Python
• Experience with cloud computing platforms such as GCP, AWS, or Azure
• Demonstrated continual learning around Artificial Intelligence
• Ability to work from home and at least three days per week in the office
• Scheduled weekly hours: 40

Core Competencies

Demonstrates expertise in building and operationalizing AI red team frameworks and conducting adversarial assessments on AI systems. Proficient in cybersecurity principles, AI model architectures, and utilizing cloud computing platforms for testing and experimentation.

Highest-signal resume keywords

• AI Red Team Framework Development
• Adversarial Assessment Experience
• Cybersecurity Principles Understanding
• Python Programming
• Cloud Computing Platforms Experience

ATS Optimization Keywords

Hard Skills

• AI Evaluations
• Adversarial Prompting Strategies
• Data Poisoning Techniques
• AI Model Architecture Knowledge
• Web Application Testing
• Local LLM Prototyping
• Generative AI Training
• Automated Testing Tools
• Risk Assessment
• Remediation Guidance

Soft Skills

• Presentation Skills
• Collaboration
• Continuous Learning

Industry Keywords

• AI Safety
• AI Security
• Generative AI
• Adversarial Machine Learning
• Cybersecurity

Tools & Technologies

• MITRE ATLAS
• OWASP
• GCP
• AWS
• Azure
• Ollama
• LlamaFiles