Monday 15 July 2019

Is JavaScript a good choice for machine learning?

First, JavaScript is not known to be a great programming language for Machine Learning. The most popular one is Python. There are great Python libraries for Machine Learning - Scikit Learn, Tensorflow, Keras, etc. Ideally, you should use Python to save your time.
Second, while learning ML, or for that matter, any concept, your focus should be on the core idea - how it works and most importantly, why it works? Implementation in JavaScript or Python or any other programming language is just a matter of writing working code. The bottleneck is generally not on the coding piece of it. The bottleneck is on understanding the concept.
In fact, a lot of people focus on the programming language, tools, frameworks, and what not. While the implementation is one of the pieces of the puzzle, it is far from being the crucial piece. The most important part is to understand the core concept and be able to solve problems using that concept.
You can have a resume full of various programming languages and tools. However, if you are unable to come up with solutions to the interview problems thrown at you, your knowledge of 10 such programming languages is completely useless.
As an exception, some companies working in a very specific domain indeed require you to be well versed with a particular programming language. For instance, all the high-frequency trading firms like Tower Research, APT Portfolio, Alphagrep, etc. expect you to have a firm grip on C/C++ programming since these companies almost exclusively use C/C++. Also, certain important concepts of C/C++ that are specific are often used in HFT firms. That’s why they prefer those with a deep knowledge of these programming languages.
But, by-and-large, your focus should be on learning problem-solving and building concepts than on learning 10 different programming languages.

2 comments:

  1. Hi Aman, this is Surya Teja.

    Our lecturers in college completely focused on Java for programming basics as well as DSA. Now , they've jumped on to Python, since I belong to the Artificial Intelligence Engineering (newly launched in the campus) branch, and we are made to implement all the ML concepts that we are being taught in Python.

    So do you recommend me to learn C/C++ side by side as well , as you've told that tech giants are focusing a lot on C/C++

    ReplyDelete
    Replies
    1. Any programming language is fine. What's more important is that you're able to explain the concept and solve the problem.

      Delete