A Tribute to Dimitri Bertsekas

This article is a personal tribute by Ben Recht to his teacher and friend Dimitri Bertsekas, who recently passed away. It highlights Bertsekas's foundational work in mathematical optimization, including stochastic gradient descent, convex optimization, dynamic programming, and reinforcement learning. The author discusses Bertsekas's underrated impact on reinforcement learning, his prolific writing and self-publishing through Athena Scientific, and his role as a passionate communicator. The piece also touches on the tension between theory and practice in the current tech landscape, using Bertsekas's work as a lens. While deeply personal, it remains focused on technical contributions and the tech industry, making it relevant to IT/TECHNOLOGY.