Master's/PHD for Data Science Job

[oli_action]

Hey everyone,

I recently spoke to a friend of the family who is a lead developer(45). I told him I am really interested in data science. He looks up at me and says:

well, you better get a master's degree then. Everyone I work with in data science has a master's and some even a PhD.(he has a master's degree)

I'm a first year CS student and have a long way to go, but I'm already thinking ahead . I want to jump right in after my 3rd year. Is it really necessary to have a master's degree to get a good, well paid data science job? And is it an advantage when it comes to getting promoted? I know a lot of programmers out there don't even have a uni degree, but are still great programmers and get good jobs( not sure if in data science though.?).
Is it all about your skills or is the "paper" still very relevant? Especially in the data science scene with all the experienced competition out there?

All the best. Oli

[Larz60+]

I think a masters degree is good to have.
You will get the better jobs if you have one, but it's also possible to land a good job with a bachelor's.
In many cases, a masters will land the job before a PHD will.
But if you are planning to work for the Government, or academia then PHD is the way to go.

I should also add that experience is a huge plus.

[micseydel]

I know a lot of programmers out there don't even have a uni degree, but are still great programmers and get good jobs

I work in silicon valley, having worked at LinkedIn and a couple of startups. In my experience, software engineers usually do have a degree, even if it's in something unrelated (e.g. English, Philosophy, etc.). I know that not everyone does, but it seems they're getting phased out over time, as competition increases; this is exacerbated by the fact that many highly skilled engineers / data scientists are coming here on visas, which are easier to obtain if you have a masters. Consider, if you're choosing between hiring two equal candidates, and one has a degree, who are going to pick?

Is it really necessary to have a master's degree to get a good, well paid data science job?

Necessary? No. I know a backend engineer at my current company who became a data scientist, with just a BA (in philosophy). That said, he made an internal change at the company after having worked there for around four years.

Just like with SWE (software engineering), as a data scientist you'll be competing with the broad market, most of which have degrees. I'd actually argue on top of this that an education in data science (or machine learning or similar) is much more valuable to data science work than computer science is for software engineers (I would estimate 85-99% of my work is based on experience, not school).

I'm a first year CS student and have a long way to go, but I'm already thinking ahead . I want to jump right in after my 3rd year.

Being so early in your career definitely gives you a huge advantage. In my experience, data science internships are more likely to allow someone who doesn't already have the degree. Such experience may offset not getting that degree, but that aside, you'll get experience in it prior to starting fulltime work and potentially getting stuck doing something you don't enjoy. I strongly recommend you seek an internship as soon as possible. (I personally have very little data science experience myself, but I know some people excitedly jump into machine learning and are disheartened when it turns out more of their time is spent randomly tinkering with variables, instead of something more cerebral.) To prep for this, I recommend you find a MOOC to self-educate, if you can't take a uni course on the topic.

And is it an advantage when it comes to getting promoted?

That will depend on the company. Some will just look at job performance for such things, and others will look at the whole person. (Again, it comes down to resourcing - if the number of promos is limited, and you're choosing between two otherwise equal people, you'll probably choose the one with the degree.)

Is it all about your skills or is the "paper" still very relevant? Especially in the data science scene with all the experienced competition out there?

Paper still matters, unfortunately, since there's no (cheap) way to measure skills without it. The paper is more relevant to getting an interview though, and once you have the interview they'll likely be evaluating you much more on that than the paper. But, the paper is a filter.

I will say, data science seems pretty impacted right now, meaning there are far more applicants than people qualified to do the work. It's hard to standout.

Lastly, I wanted to share my personal experience - I was interested in data science for a long time, and in 2017 I was offered a data engineering job, in spite of having been a backend engineer prior to that. (I currently do both.) What I've learned from that experience is that even though I do enjoy playing with data, that data scientists work is not what I'm interested in. I hope I don't offend anyone by saying this, but in my experience data scientists write very ugly, difficult-to-maintain code compared to engineers (this has been true at every company I've worked at, though I saw far less at previous companies). Also this will depend on the company, but data science may involve a lot of SQL, and while SQL is great in a lot of ways, I've had to debug queries that have dozens of joins and are dozens of lines long and it's not fun.

As a kind of aside, I want to mention one more thing. I took the data engineering position because I thought it would help me pivot to data science. I think it could have, but I chose to stay where I am. You might want to consider data engineering if you're interested in doing adjacent work, and (in my case) that gaining you access to the data so that you can play with it if you want to. I have a task (approved by my manager) to look for Benford's law in our company data, and if I find anything interesting to write a post for our tech blog. So I'll reiterate - where you're currently at in your career, you have a lot of opportunity to experiment and not get locked into anything, so definitely try stuff out!

[micseydel]

For what it's worth:

I don't know Vin, but Mikiko is a data scientist at my company. I haven't worked with her much, but she made a good first impression.

[lolapaluuza]

I met too unreasonably high requirements for some vacancies for which they then took dudes who worked with interesting cases but did not meet 90% of the resume requirements.