Kaivalya Hariharan

Kaivalya Hariharan is an MEng AI candidate at the Massachusetts Institute of Technology and the founder of the MIT AI Alignment group, based in Cambridge, Massachusetts. He earned his bachelor's degree in computer science and mathematics from MIT in 2024, where he served as a SuperUROP scholar investigating neural network failure modes during his undergraduate studies. Hariharan currently conducts research within the MIT CSAIL LINGO group under Jacob Andreas, building upon previous artificial intelligence research stints at Redwood Research and Truthful AI. His academic contributions include a 2022 Best Paper award at the NeurIPS ML Safety workshop, followed by multiple publications at the 2023 and 2024 NeurIPS conferences, alongside his selection as a 2025 Emergent Ventures Fellow. His ongoing work focuses on advancing machine learning safety, neural network evaluation, and the alignment of large language models.