Is a MS/MCS degree necessary to work in machine learning engineering

J.F. - good question. It's great that you're thinking ahead and looking at job descriptions online now. I would use what you're finding as the best answer to your question: if 75% of the programming positions you're interested in actually require (or even prefer) Master's, then that should tell you exactly how important it is by qualifying you for that many jobs!

From a compensation and career trajectory standpoint, I would strongly recommend at least finding a reputable 1-yr Master's program even though it's not "necessary". Not only will you give yourself an advantage over competing candidates when you apply for jobs, but you'll have a better opportunity to build a more diverse coding repository for your resume, and you'll be more effective when you start coding professionally.

As for the MCS vs. MS, I can't speak to how much that matters but my sense is that most positions you're looking at would be OK with either. I'd imagine that by the time you're applying for your 2nd job, whenever that is, your professional experience will matter more to whomever is reviewing your resume anyway. Though, when applying for Master's programs you might want to look at the alumni of the different programs and see if there is a trend among them that helps you identify your ideal path.

It is not. It is helpful for sure as most studies, as academic experience is. However far more important in my experience are the projects you have been part, the code you have written, a good repo is always welcome. Some hiring managers and hr people will have valid reasons to hire only Ph.D' or masters, however as a whole the industry is looking for people that know how to do stuff, understand how it works and go and do it.

Than being said, it is highly advisable that:

• You have good foundations on coding, be proficient (good) in at least one proficient in one programming language, feel at ease using it. You don't get to say I would Google it when you are writing a for loop, we all use Google as our API to Stackoverflow, but the simpler your queries the more evident you don't know your stuff.
• Understand how machine learning works, you don't get to say you know machine learning because you have plugged a dataset to the sklearn API. You need to understand the data, the algorithm, the metrics, the validations, runtimes, big O and memory concerns, all them are important.

So no, it is not necessary as long as you know what you are doing and can prove (show) it.

I have a slightly different experience that I'd like to share with you - I got a MS in Computer Science and most of my batch specialized in Machine Learning since its a super popular skill right now. However, majority of them ended up in a general software engineering role and did not end up working on Machine Learning. While a lot of people will recommend having a MS or a PhD degree, it won't necessarily guarantee you a job in this domain even if you specialized in it.

An alternate strategy that I have seen a lot of people use is to move into Machine Learning at their existing jobs. A lot of companies these days require to do some kind of Machine Learning and they are usually more than happy to support engineers who are keen to take on new roles in the company. That might be an interesting strategy that you could potentially explore.

While a Masters or even doctorate is not normally required to work in software development like Artificial Intelligence (AI) or Machine Learning (ML), any additional Certifications and Educational Degrees that you obtain will make you more competitive in the job market in general. There are many nondegree training programs and certifications that will prepare you for a career in software development.