Quis custodiet ipsos custodes? (but for guardrails)

Hi, I’m Kevin. I’m a technologist that works on AI strategy, with particular attention to how frontier capabilities are reshaping human cognitive agency. My technical research looks into latent model biases and how AI can manipulate human decisionmaking. I’m also navigating this world on the security end with how AI is reshaping Pacific geopolitics and how the U.S. is approaching this grave new world. I’m a Master in Public Policy candidate at the Yale Jackson School of Global Affairs, affiliated with the Schmidt Program on Artificial Intelligence, Emerging Technologies, and National Power, and I serve as Editor-in-Chief of the Yale Journal of International Affairs.

I kind of straddled both worlds in AI policy between the people who build frontier systems and the people who write rules for them. I went to graduate school to pivot to policy from the technical side, after several years as a machine learning engineer and data scientist in the U.S. national security community, and I try to write for readers on both sides of that divide.

→ The Research page sets out the three streams of my current work.

→ The Writing page collects my public writing as it lands.

→ The Speaking page tracks talks, panels, and guest lectures.

→ The Teaching page covers the Yale courses I help teach.

What am I up to now?

• At the Centre for the Governance of AI, I’m a Summer Research Fellow measuring how reliably auditor models detect misaligned behavior as interaction horizons lengthen, benchmarked across frontier and open-weight architectures. This is currently being fine tuned for a NeurIPS workshop submission.
• I’m co-authoring a policy paper with researchers at OpenAI on global frameworks for advanced AI access and compute governance.
• At the Council on Foreign Relations, I’m a graduate research assistant working on recursive self-improvement and international security with Adjunct Senior Fellow Gordon Goldstein.
• With Doublethink Lab, I study foreign information manipulation and interference in the Taiwan Strait, including an October data sprint in Taipei.
• With the California Governor’s Office of Emergency Services, I’m translating the state’s SB 53 frontier AI framework into public-sector crisis management protocols and intervention thresholds.
• I’m co-mentoring a SPAR project this fall, Assessing Taiwan’s Global Position for Transformative AI, with David Sanchez Garcia which is open to mentee applications through 18 August.

What did I do before this?

• I spent several years as a data scientist and machine learning engineer working on AI in the U.S. intelligence community.
• I hold a B.A. with Quantitative Social Science from Dartmouth College.
• I straddle geopolitical policy advising and full-stack technical work on AI. I work in Python, R, and SQL. I can also read Chinese and speak Mandarin fluently.