Confidential
Hybrid (Flexible / Intentional)
Belgium
Posted 4 days ago
Company: [Your Company Name]
Location: [Remote / Hybrid / City, France]
Employment Type: Full-time
Department: AI Ethics & Cybersecurity
About the Role
We are seeking an AI Safety & Red Teaming Analyst to stress-test our AI applications, discover vulnerabilities, and ensure our systems operate safely, ethically, and securely. You will act as an internal adversary to identify edge-case failures, prompt jailbreaks, and potential compliance risks before models reach production.
Key Responsibilities
- Red Teaming & Security Testing: Design and execute adversarial testing scenarios to identify prompt injection, jailbreaking, and data extraction vulnerabilities.
- Bias & Fairness Audits: Evaluate model outputs for implicit bias, misinformation, and ethical risks across diverse demographics and use cases.
- Guardrail Development: Work alongside engineering teams to implement content filters, safety layers, and automated moderation pipelines.
- Compliance: Ensure model deployments align with regional and global regulations (e.g., EU AI Act, NIST AI Risk Management Framework).
Qualifications & Requirements
- 2+ years of experience in cybersecurity, red teaming, ethical hacking, or AI safety research.
- Deep understanding of Generative AI vulnerabilities and common attack vectors.
- Familiarity with Python, security evaluation tools, and automation frameworks.
- Strong interest in AI ethics, governance, and risk mitigation.
Offer
- Competitive compensation package with equity options.
- Dedicated focus on cutting-edge security and ethical tech deployment.
- Collaborative and inclusive company culture.