Fractal is looking to proactively identify, simulate, and mitigate threats to AI systems across UHG, shaping the strategy for adversarial and non-adversarial testing of LLM models to ensure their robustness, fairness, and security.
Requirements
- 10+ years of experience in Machine Learning
- 6+ years of experience in cybersecurity, red teaming, or adversarial testing
- 3+ years of experience with generative AI systems (e.g., LLMs, agents) and their unique threat surfaces.
- Familiarity with AI safety, fairness, and interpretability frameworks.
- Contributions to open-source tools, academic research, or AI security communities.
- Knowledge of secure model deployment practices in cloud and edge environments.
Responsibilities
- Define and execute the vision and roadmap for AI red teaming initiatives across the enterprise.
- Build, mentor, and lead a high-performing team of AI red teamers, adversarial ML researchers, and security engineers.
- Oversee the design and execution of red teaming exercises targeting AI systems
- Partner with AI/ML engineering, enterprise cybersecurity, product, legal, and compliance teams to embed red teaming insights into model development and deployment lifecycles.
- Ensure alignment with internal policies for Responsible use of AI
- Stay ahead of the curve on adversarial ML, model exploitation techniques, and AI safety research; foster a culture of continuous learning and innovation.
- Lead investigations into AI-related security incidents and develop mitigation strategies and post-mortem analyses.
Other
- 6+ years of leadership experience managing technical teams and cross-functional initiatives.
- Strong communication skills with the ability to influence executive stakeholders and translate technical findings into business impact.
- Experience working in regulated industries (e.g., finance, healthcare, defense) with AI governance requirements.
- Advanced degree (MS or PhD) in Computer Science, Machine Learning, Cybersecurity, or a related field.
