I will share some experience on writing a software to play the game of Go in Kotlin. The game of Go was born in China more than 2000 years ago and despite its very simple rules is considered more complex than chess and so one of the most interesting from a software developer point of view. Until 2018 when AlphaGo beat the world champion, it was considered humanity’s last stand. I’ll present a didactic Go bot engine I developed in Kotlin. It’s very simple but it is already able to give interesting matches. The session will be split in three parts:

First, an introduction on the game of Go and why is so interesting.
The second part is about some game algorithms, side to side with Python equivalents to compare the two styles.
The third part includes some considerations on performance, how to identify bottlenecks, and how to improve them, and some counterintuitive facts I discovered in the meanwhile. As a final note, I’ll offer some insight on current work on a deep learning engine to bring it to the next level. All the code is available on GitHub https://github.com/uberto/kakomu