STAIR

Dhruv is a sophomore at Stanford University pursuing degrees in mathematics and computer science, working with Rylan Schaeffer and Andres Carranza. He primarily focuses on mechanistic interpretability, learning dynamics, and the “deconstruction/reconstruction” framework proposed by Rylan. He particularly enjoys working with toy models of poorly understood emergent phenomena. Dhruv has previously worked in various machine learning labs at Harvard, NASA Harvest, and Regeneron Pharmaceuticals. Dhruv’s hobbies include growing mushrooms, arm wrestling, and cats. He loves to talk about cutting-edge science in bioengineering, computer science, material science, and physical systems, so feel free to reach out and schedule a cofffee chat!

Search for Dhruv Pai's papers on the Research page