San Francisco, USA

Athan Reines

Full-stack Engineer, Stdlib

Athan is a full-stack engineer and data scientist. He has a PhD in Physics, where he used machine learning and time series analytics to probe biological systems at the nanoscale. He currently works full-time on open source projects to facilitate numeric computing in Node.js and JavaScript. For his latest open source project, see stdlib, a standard library for Node.js and Javascript.

details about the talk

Node.js Addons for High Performance Numeric Computing

Node.js add-ons allow native code to be run from the Node.js runtime. In this talk, Athan will discuss how to utilize native add-ons for high performance numeric computing and machine learning in Node.js applications. He will first provide an overview of add-ons and their associated toolchain. Next, he will walk through an example which involves compiling basic linear algebra subroutines (BLAS), a suite of libraries which are a core foundation of most modern numeric computing environments, as native add-ons. While Node.js add-ons are oriented toward C and C++, he will show how to extend compilation support to Fortran libraries in order to maximize computational performance. Throughout the talk, Athan will offer lessons learned and other insights gained while writing add-ons and demonstrate why Node.js is an excellent environment for high performance numeric computing and machine learning.