Just changed my major to computer science and I intend on specializing in machine learning...

Depends on how much you want to put in.

If you are doing your CS degree, scoring well and that‘s it, then it is literally useless. Its just gonna be standard algorithms, some maths, some coding. Like EVERY other major, all that matters is what you do in your free time.

If you grind kaggles, build image classifiers, and maybe even screw around with the OpenAi gym, then you could land a very good job.( I am in Europe, but it probably isnt that different)

If you roll up to a job interview and just spew some stupid bullshit about logistic regression with no github to back it up, they will laugh you out of there. If you show up with crisp projects using Tensorflow/Keras/etc., that use all the latest and coolest models and do something that boomers find amazing( «Oh wow your neural net can tell whether there is a child in the picture xD»), then it will be good. Plus, since youre in the U.S, you may eventually make your way into a startup

Career options will be varied, like anything else this will likely be determined by your internships.

Finance- H.R guys in finance love ML. They see all these quantitative hedge funds and wanna be like them, you just have to position yourself as their ticket. Just meme about learning how something predicts stock prices and youre gucci.

Google/Facebook/etc-You can imagine why, but the interviews will (in my estimation) be difficult. In Europe, I don't even think you can get into a Facebook ML role without at least a masters, more likely a PhD.

Startups-Yh whatever the fuck they are building they will want to sell to investors using "muh AI".

Just make sure you are actually knowledgeable. You most likely won't be able to compete for the Google/Citadel/Facebook jobs with the Harvard/CarnegieMellon/MIT guys, but you could get the remainder of the pie quite easily, especially in Utah.

>specialising in ML
That’s a pretty shallow specialisation and you’ll have it all covered in a couple months. Can’t you think of anything else?

>What kind of study can I expect to encounter?
Data, data and more data.

>i want to specialize in outsourcing my own specialization

Yh, but the kicker is that automation is taking longer than expected and its still a big tech arms race, with an insane shortage of real skill.

You just look at the research departments of facebook and google, you realize youre getting paid 200k+ for doing research you would do in Uni.

If you get in right now, you could make more in 10 years than most jobs do in 20

nice one pal you got in right at the end as programming related salaries end up going down to the average due to market saturation

If you want to do anything interesting (and not in a shitty startup) in ML, you need a PhD. Just look at job postings at Google etc.

>I intend on specializing in machine learning
How are your math skills, user? I've spent around 3 years in my current 'Intelligent Systems' course - basically, it's comp. sci. with a broader focus on ML theory, application and robotics - interesting stuff btw.

When the course first started, we had around 60 students and now on the third year we are down to 35 students (including me). The students who left the course either transferred to just plain comp. sci. or dropped out entirely from uni. The main reason on why they left the course was that they couldn't understand it (and were struggling with the course) or just took the course because of the buzzwords that get thrown around in media (like big data, AI, ML, self-learning, automation etc.) and how every slag from HR would suck dick to get someone who has an 'ML degree'.

Anyway, as for the studying you will probably need to brush up on your maths skills since American maths education is a joke when compared to Europe. Also, don't listen to Jow Forums - most of them who like to meme 'AI/ML' have no understanding of what it is and are just a bunch of pathetic linux systems admins (at best) who were bullied in school.

(pic unrelated)

Attached: 1564599549606.png (285x313, 228K)

learn your math, that separates the "coder" from a real computer scientist

>algebra and calculus are hard
nice try europeon

I work in machine learning. DONT MAKE MY MISTAKE!

Since it is the machine that is doing the learning, you only have to provide it with data. So you are just programming data loading and file formats.
I shouldn't have to explain how boring that is.

Additionally, the scientific community around ML only excels in arrogance. They're not the smartest, but they think they are.

I hope to be back to a real engineering job in a few months. ML is sheit.

>linux systems admins (at best) who were bullied in school.
t-take it back

Attached: starve.png (640x360, 172K)

Why is ML research held in such high regard then? Everyone acts like it's not only the most difficult subfield, but also the most prestigious.

>doesn't know how to program
>doesn't know math
>already chosen specialization
very nice

>waah I can do basic level maths but can't apply these same skills to any useful scenario
okay retard

machine learning is just the parallel and/or hierarchical application of bayesian methods
I think you will learn matrix math and nothing else since 'ML' people just glue together a bunch of theano or tensorflow classes.

It's up there in the buzzword bullshit with blockchain, AI, and syntactic sugar.

Like cancer research, there's a lot of money to be milked

Take a double major in math. Take as many math electives as you can and learn how to prove shit. Take some analysis courses also so you're not a complete brainlet.