Solving the Hard Problems: Combining AI and Physics for PCB Design

In this fascinating interview, Sergiy Nesterenko details the journey that ultimately drove him to solve a critical bottleneck in hardware engineering: the manual, slow process of printed circuit board (PCB) layout.

In his discussion with our Moore's Lobby host, Daniel Bogdanoff, Sergiy shares engaging anecdotes from his early career, including a SpaceX internship where he was assigned to literally watch paint dry. He also describes his first attempt at PCB layout at SpaceX, where a power switch board caught fire in his hands. This experience fueled his desire to bring robust automation to the PCB design field.

Sergiy explains why traditional auto-routers have failed to gain widespread adoption for sixty years. He then explains how Quilter solves this by leveraging reinforcement learning and computational physics.

They discuss several major technical insights and future directions, including:

-The software compiler analogy.

-Physics-based automation.

-Tackling complex boards.

-His 10-year vision.