About Me
I’m a Computer Science Ph.D. Candidate at the Stanford Trustworthy AI Research (STAIR) Lab and the Stanford Intelligent Systems Lab (SISL), advised by Prof. Mykel Kochenderfer and Prof. Sanmi Koyejo. My research sits at the intersection of AI and policy:
AI systems are increasingly used in high-stakes settings, from medicine to hiring to national security, and the evaluations we run on them shape deployment decisions, regulatory obligations, and, if used in training, the model behavior itself. If those evaluations are unreliable or measure the wrong things, every decision built on top of them is on shaky ground. So my research asks a seemingly easy question: do AI evaluations reliably measure what we think they measure? I work on this across three layers. First, the science of evaluation itself: what makes an evaluation valid, reliable, and worth trusting. Second, the ecosystem around evaluations: how results get reported, aggregated, and used, and what incentives that creates for developers. Third, the governance layer: where institutions need technical capacity to act on evaluation evidence, and how to build it. I am grateful to have received one of Stanford’s Interdisciplinary Graduate Fellowships for my research.
Outside of academia, I’m an avid traveler (60 countries and counting), a triathlete, audiobook listener and PS5/Switch 2 player. I love meeting new people over a cup of tea or a walk outside. Feel free to reach out at anka [at] cs.stanford.edu or DM me on X/Twitter (@AnkaReuel).
I also have a sweet tooth for crazy out-of-the-box ideas and adventures. In 2018, I drove a mini motorcycle around the Gates of Hell in Turkmenistan. How I got there, you ask? Long story that I am happy to tell, but essentially I took an Opel Corsa that I bought on eBay for 300 € and drove it from Germany to Mongolia and back. I told you, I like crazy.