I found an improved algorithm for unitary synthesis! This is a fundamental step in compiling quantum programs, and pushes state-of-the-art closer towards the theoretical lower bound. This problem has been on my mind for the past 2 years and I'm happy to make progress.
