2
$\begingroup$

I see there's a lot of machine learning job openings with skills requirements, python ,R, keras,tensorflow,pytorch,spark, etc.which are completely fine & reasonable, but why many of the recruiters include C++ ,like what is use of C++ in ML research, or even creating ML pipelines ?

How much C++ should I know if I'm good at rest of ML skills ?

$\endgroup$
  • $\begingroup$ Sometimes it's simply because the company has some of its code base in C++. $\endgroup$ – Erwan Aug 5 at 22:32
1
$\begingroup$

C++ is often listed not because you will necessarily be coding in C++, but rather fundamental understanding of memory allocations, object-oriented design patterns, and other CS fundamentals that python (and other similar) languages abstract a bit.

Plus, there's Cython and other c-derivative frameworks that may be necessary to understand in a real-time environment.

$\endgroup$
3
$\begingroup$

Many machine learning products emphasize on speed over accuracy, e.g. in autonomous vehicles, a model that predicts in milliseconds with 98% accuracy is more valuable than a model that predicts in seconds with 99.99% accuracy.

While Python provides many great packages for exploring machine learning research and testing proofs of concept, it is significantly slower than C++ and not acceptable for products with real-time requirements. If you want to know why Python is slower, check out this thread.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.