Sunishchal Dev is a Technology & Security Policy Fellow at the RAND Corporation, where he evaluates and governs AI systems with a focus on biological and chemical threats.
He is a lead author on the RAND report "Toward Comprehensive Benchmarking of the Biological Knowledge of Frontier Large Language Models," which evaluates 39 of the most capable AI models against biological and chemical knowledge benchmarks. The research found that frontier LLMs, led by reasoning models, are exceeding expert human performance on biology laboratory protocol and graduate-level question-answering benchmarks.
Prior to joining RAND, Sunishchal received a B.A. in Technology and Innovation Management from University of Washington, Bothell. He has a decade of industry experience as a data scientist and management consultant implementing AI solutions for enterprises.
As a MATS Scholar, he studied AI-risk mitigation and helped develop the first publicly available curriculum for AI safety evaluations. He learned how to train a GPT from scratch, interpret neural networks, and use reinforcement learning.
At Algoverse, Sunishchal has been mentoring students and professionals on novel AI research projects, including examining the reliability of AI safety benchmarks. He authored an ICML main conference Spotlight paper while at Algoverse.
Begin Your Journey
The application takes 10 minutes and is reviewed on a rolling basis. We look for strong technical signal—projects, coursework, or competition results—and a genuine curiosity to do real research.
If admitted, you will join a structured pipeline with direct mentorship to take your work from ideation to top conference submission at venues like NeurIPS, ACL, and EMNLP.
