Fazl Barez
AI Safety and Interpretability Researcher
With extensive experience in AI research and safety, I help organizations develop and implement responsible AI systems. My work spans academic research, industry applications, and policy development, with a focus on making AI systems more interpretable, reliable, and aligned with human values.
Areas of Expertise
- AI Safety Assessment - Evaluating AI systems for potential risks, failure modes, and alignment issues using state-of-the-art methodologies
- Interpretability Solutions - Developing techniques to understand and explain AI system behavior, with particular expertise in large language models
- Safety-First Architecture - Designing ML systems with built-in safety mechanisms, transparency, and robust evaluation frameworks
- Technical Standards & Policy - Advising on AI governance, safety standards, and responsible development practices
I contribute to the AI safety community as a keynote speaker and panelist at major conferences including NeurIPS, ICLR, ACL, and ICML. I serve as a programme chair, reviewer, and workshop organizer, while also mentoring PhD, MSc, and BSc students in AI safety research.