About Me
My professional background is rooted in facilitating cross-cultural understanding, supported by graduate training in Middle East Studies and Arabic. That foundation informs how I approach complex, high-stakes decision-making across diverse perspectives.
I entered the tech field at Facebook, where I spent three years building and scaling operational programs in escalations and user support that affected billions of users. This work taught me how to design and pressure-test systems with real consequences at scale.
I later transitioned into Responsible AI at Google, where I spent six years shaping fairness policies and evaluation approaches and translating sociotechnical research into practice. The work spanned advising senior leadership, partnering with research, policy, and product teams, and leading applied AI ethics programs that built shared understanding across students, employees, and executives while connecting research insights to real organizational decisions across multiple product areas.
Currently, I serve as Director of Responsible AI at Salesforce. In this role, I lead evaluation framework development and ethics review processes that embed accountability and ethical decision-making into how AI systems are built and deployed.
Across roles, my work has always been about translation: turning complex principles into actionable systems that function in practice.
Education
MPS, Arabic | University of Maryland, College Park
MA, Middle East and North African Studies | University of Michigan, Ann Arbor
BA, Middle East Studies and Spanish | Wellesley College
This site is where I share lessons from the field. I write about what has worked, what hasn’t, and how organizations can build Responsible AI governance systems that are credible, durable, and aligned with how products are actually developed.
My Operationalizing Responsible AI Principles
1. Translate principles into decisions
The real challenge of Responsible AI is turning ethical principles into actionable decisions that teams can make under real-world constraints.
2. Embed Responsible AI into workflows
Responsible AI is most effective when it is built into product development and review processes, not when approached as a separate or purely theoretical exercise.
3. Build practical governance systems
Organizations that succeed integrate governance, evaluation, and accountability into product workflows, creating scalable, durable systems rather than relying on performative compliance.