Is university the best route for a career in machine learning?

Are online courses really worth it?

I noticed that software development has taken huge strides in the last 10 years to push for self-taught developers, apprenticeships and intensive code camps, or online short course providers like Udemy, Udacity and Coursera. But are these really a substitute for university?

Ironically, I’ve spoken to a few hiring managers who are totally put off by seeing these courses on a candidate’s CV. They think that if the engineer was a capable one, they wouldn’t need to supplement their learning with a short course, and it suggests they’re not ready to design, develop and ship machine learning models in a fast-paced environment.

The university route

Many clients I’m now working with in the UK desire a masters degree as a minimum. They think this will give a person a solid understanding of the software engineering theory and fundamentals they will need if they’re to build robust and scalable models in the real world.

However, I am seeing a shift with some companies being more accommodating and open-minded when it comes to education.

For instance, a London-based AI start-up is willing to interview candidates on the strength of their GitHub repository, and a machine learning task at interview stage.

This, for me, is a step in the right direction as some people just don’t have the access to higher education. This means candidates will now get the chance to prove themselves and be judged on the quality of their code.