Architecture of math calculations on Kotlin

 /  RU

During this session, we'll talk about KMath (https://github.com/mipt-npm/kmath) and more. With Aleksandr, we'll consider different approaches to math API and their realization in different programming languages (Python, C++, Julia, Java and Kotlin). We'll talk about why it's hard to make math both convenient and fast. And take a closer look at the boxing problem.

The audience will see how the context-based approach in Kotlin helps solve not just the balancing problem of speed and convenience but also allows you to make math libraries modular and ensure their compatibility with high-performance platform libraries.

In conclusion, let's talk about the future of math libraries outside C++ (and only in Kotlin).


Speakers

Aleksander Nozik
JetBrains Research

Aleksandr teaches physics, mathematical statistics, and Kotlin at MIPT, he's the deputy head of the nuclear physics methods laboratory at MIPT and the head of the group of the same name at JetBrains Research.

He also has a Ph.D. in particle physics, and more than 12 years of development experience, including commercial development. Mostly in Java, but the last 4 years in Kotlin, apart from that there have been Python, Groovy, Julia, and so on.