Is Machine Learning/Deep Learning the ultimate career path?

>>62140073
No, it's only good for solvable problems or poor approximations.

>>62140073
it's a meme. just like big data. it's just linear algebra done over and over again

Nah, it's a meme

>>62140107
let me clarify, deep learning is a meme. machine learning is a pretty solid career path

>>62140107

>>62140093

>>62140108

Would you say, Distributed Systems is a safer bet?

>>62140107
>applied regression and stochastic analysis is linear algebra
It's time to stop posting.

>>62140134
don't invest too hard into a niche. i.e. don't get a degree just for ML or just for distributed systems unless that's all you plan to work on for the rest of your life. get a masters in CS and then just do a specialization. it gives you a lot more options

>>62140134
No, it's a bubble. You should just become a carpenter and sell handcrafted furniture to middle class goons who think they're rich.

>>62140187
My anxiety arises from the fact that recruiters are mostly dicks who look for PhDs and the fact I'm not a child genius in math nor I see myself reading volumes about stochastic math

>>62140073
Companies are still hiring for devs using 30+ year old technology. Not concerned even if you're correct.

>>62140222
Lol, or join a pseudo cult. We don't have scientologists in my country, but there are plenty of hippies going "The Art of Living Way?'

What is the CS field that /g/ aproves?
[spoiler]inb4 webdev[/spoiler]

>>62140239
Undergrad was in Computer Engineering/C.S w/a focus on embdded sytems.
Grad was distributed systems

We covered what is known as AI under (constraint optimization). It was boring and dry as is the industrial application. Dont waste your time. If you want to specialize in grad school try out computre security or focus on Distributed Systems (broader reach). You can dip into AI under Distributed systems.

>>62140278
> Embedded Systems
> Computer Security
> OS

>>62140271
Revolutions usually happen suddenly and are usually brutal (sorry, not wanting to sound like Che Guevara, my father has a PhD in History). Point being, that could change fast. Plus, you don't want to work in a 30 year old technology

>>62140312
I'm currently working for an AI+Security Startup, in Scala. Thing is I'm not loving the data science part (altough I do like Scala)

>>62140374
The Datascientis earn BIG MONEY though, hence my question about obsolesence

Bump?

File: h4sH5IB.jpg (21KB, 507x338px) Image search: [Google]

21KB, 507x338px

>>62140332
I have MSc in embedded systems but couldn't find a job, then I became a full stack code monkey

>>62140239
You don't need to be a genius to get a PhD, but you shouldn't do it or later work in a sphere that you don't enjoy at all. ML is moving very fast right now, so if you want actually be useful you pretty much have to live and breathe it. Data science is nice moneywise, but these are phenomenally boring jobs that require a great amount of knowledge - basically a slightly more techy actuary. Unless you're in some hot startup or fintech/QA 80% of your day will be cleaning data instead of being a ""scientist"". Moreover the term "data science" itself has been memed so hard lately, that every two-bit statistician is one nowadays, so it's pretty hard to differentiate between "true" vacancies and those with fancy title. As other anons said, unless you're very interested in these particular topics you'd be better off specializing in something more broadly applicable. Your goal should be gathering as much different experience and contacts during early career, because in 5-7 years you'll inevitably want to move on to a more managerial profile anyways.

>>62140073
AI in general is pretty solid but it's nothing like in the movies.
It's just numerical analysis with some linear algebra sprinkled over. Basically go for it if you like to watch a lot of numbers.
Neural nets can in theory approximate any computable function, but we're not getting anywhere interesting for a common person with current approaches.

>>62140579
Sounds like the kind of nightmare I have. Having to learn React, Angular, Vue or whatever the hell you frontend guys use

>>62140595
Thanks for the reply everybody. Actually, no, I don't want a managerial role, that's why I'm looking to specialize

>>62140608
I really loved embedded programming and even published a paper in it in a decent journal, but fate was so unfair to me anyway, now I have to compete with people who never got to university for the rest of my life

>>62140640
I'm 29 btw

>>62140073
How long until hardware functions like a brain? E.g. being a seemingly non-deterministic system that can, among other things, approach arbitrary objects fluently? I feel that emulating everything in software is not the solution.

>>62140645
What about going the DevOps + mix CyberSecurity buzzwords route?

>>62140668
That and Quantim Computing would be inmense revolutions, while not impossible, they render this preocupation absolutely obsolete

>>62140640
>I don't want a managerial role, that's why I'm looking to specialize
Then go for something you can become very proficient and irreplaceable in and possibly not very popular. Age discrimination is not a meme in CS/IT.

>>62140668
We don't even have approximate understanding of consciousness, so not in the next few decades.

>>62140733
I know, that's why I was thinking about Machine Learning as the ultimate career path. Is there another alternative where you can earn good money and not get replaced by a 25 year old?

>>62140768
(((they))) pushing everyone to learn about ML now, it's like "everybody should code" campaign that was 5 or 6 years ago

>>62140073
What makes you think it AI didn't replace you or the ones around you already?
Did you thought the possibility this taiwanese cartoons board is actually full of AI based bots learning from a few real humans posting?
Maybe even this entire thread was responded by multiple AI's and you didn't realize it yet.

>>62140833
Still better than getting a PhD in Woman Studies if you ask me. At least they won't be uttering "you want that with fries" for the rest of their professional career.

>>62140768
It is a nice career path, but it's very knowledge intensive, credential-dependent and as you yourself said you don't particularly enjoy the underlying science/methods. Embedded, enterprise systems, RTC, cybersec are all nice. Although the latter is probably a bit vague as it covers anything from 14 year olds who know how to config netfilter to cryptography researchers and whitehats. All in all you need a specialization where your decade-long experience will be more useful and valuable than some kid's proficiency in flavor of the year framework. Find one that you enjoy/don't hate and you'll be golden.

>>62140701
this
>>62140668
we're still a really long way from that
still probable mind you

>>62140701

Quantim Computing even if it could probably happen, it won't be for mobile or PC devices, they can make cheap quantum cloud machines so they can move much processing on servers instead of being at users

>>62140864
I'm majoring in CyberSecurity in my meme third world university. Thing is my professors are the typical "managers that installed Kali and now think they're hackers" type. So I'm Learning mostly about ISO 27001 + some owasp guidelines than actual hacking.
Thanks for the advice though

>>62140854
For being an AI you're missing a lot of intelligence

>>62141046
I'm still learning, don't be mean :(

Bumpity bump

>>62140645
Any links to your paper?

>>62140073
It's a meme. Everyone wants at least a Master's Degree and 10 years experience for those "jobs". You won't get a Coursera certificate and come out with 6 figures.

What should I learn first before getting into AI?

>>62141540
the good stuff is exceptionally math heavy

>>62141521
+1

>>62141531
I have a friend with a MsC in Machine Learning from insert high tier France University. He earns as much as me, a shitty GPA Scala Engineer

>>62140073
>Is Machine Learning/Deep Learning the ultimate career path?
imho it was ten years ago, now it's overhyped with millions doing it

also, any brainlet can run those shitty algorithms after 5 minutes of research with modern libraries, there's no skill required unless you do basically a phd in stats

>>62141614
TL;DR: we're all fucked

>>62141531
ML jobs mostly look for PhDs fresh outta research. Masters is not particularly valuable and experience less so than in other CS jobs. Also OP stated he's doing a Masters, not Coursers. Stop engaging in topics you don't understand and go away.

>>62141614
ML is developing "those shitty algorithms", my retarded friend.

>>62141540
Math up to partial differential equations, linear algebra and numerical analysis. Also throw in any statistics short of the ones that use real analysis. That should have you set to read and apply most of the material. Getting hired is a different story.

>>62141669
95% of people who are into ML are just doing dumb brutal force tasks because they alot of data
it's a meme

>>62141669
what are you gonna be developing you fucking idiot? don't confuse it with applying
there are like 5 people in the world developing genuinely new ideas / decade

>>62141614
MLP is weak AI, it's basically a glorified bunch of logistic regressions over a huge dataset.
That doesn't mean the entirety of the ML area is MLP, he'll that's old news nowadays.

>>62141717
That's something I've been told. Lots of data cleansing. It makes something like DevOps or coding in Golang sound a lot of fun

>>62141632
Coursera is a possibility though, I was told by someone working in computer vision at GoPro that Andrew Ngs specialization is pretty good (and he also has a MsC in Machine Learning)

>>62141716
>autistic screeching
Please, go back to best anime girl and vidya card threads.

>>62141540
some linear algebra
"calculus"
probability theory

that's enough for most shit

they'll learn to program themselves

>>62141824
this is like saying tat cars are gonna drive themselves in 1900s

>>62140768
There will always be new meme technology that suddenly everyone needs to learn pr get replaced by a 25 year old.
The key to having a long career is to always keep learning new things every few years and staying ahead of whatever new meme technology that comes out.

>>62141881
>work in frontend web
>have to learn 5 new technologies a year
>work in enterprise
>have to learn 5 new technologies a decade
There is some difference, anon.

>>62141925
>have to learn 5 new technologies a year
i wish
its more like
>get to learn 5 new technologies a year but you never actually get to because your company is still dead set on sticking with angular 1 for another decade

>>62141925
yeah but if you're a C++ programmer or something you never stop learning what you are forgetting

>>62140073
I work in data science and this is how it works:

>85% to 95% of the work
Figuring out how to clean, store, and retrieve data in computationally efficient ways, and then maintaining those systems

>The rest (the interesting work)
You need an MS/Ph.D in math/stats/CS to even touch.

It's an overrated field unless you have an advanced degree, imo. You also have to deal with dumbshit managers who have no idea what "machine learning" really is, but propose some problem and just say, "We can use deep learning to solve it!"

>>62141938
>dumbshit managers who have no idea what "machine learning" really is, but propose some problem and just say, "We can use deep learning to solve it!"

to be fair that's exactly what deep learning is for

>>62140107
t. AI

>>62141933
kek

>>62140645

USA?

How?

The market doesn't seem that bad for embedded work when I look

>>62141521
> anon asking me to reveal my identity
how about no?
>>62142151
I live in Egypt

>>62142177
go to germany ffs

>4% unemployment, flooded with uneducated arabs but can't satisfy the huge demand for your profession

>>62142210

how can I do this now? I am almost 29, the last time I saw c code was 3 years ago when I completed my thesis, shouldn't I get a job offer then get a visa?

>>62142248
>shouldn't I get a job offer then get a visa?
this is why the immigration system is broken

ISIS and criminals just go and pretend to be refugees, while educated people hesitate and endlessly ask stupid questions

>>62142248
They will give you the Visa if your worth, start looking for the job first

>>62142248
btw, 29 is not old at all by european standards

people in the EU regularly start/change careers in mid 30's when they finish their education

>>62142248
Search for employers, send resume, do Skype interviews, obtain job offer, get a blue card.

>>62142210
Not him. What are the most requested professions in Germany? Also, with the refugees and all that, is it worth it?

>>62142248
Just go there and say you're Syrian.

>>62142654
>s it worth it?
If you're from Egypt what do you care

Probably in the distant future we (computer workers) will be outmoded, but that's not the right way to approach it.

Take, for example, the story of my boss from my first non-mcjob.

>Graduated high school in 1980
>Immediately went to work for high school as one of two A/V + telephone techs
>Always moving forward, implemented a (cool for the time) TV studio + system in the schools
>1990's rolled around, learned how to work with fiber + data networks
>Learned maintenance and repair of digital projectors
>Implemented VOIP in 2000s
>2010s rolled around, learned and implemented IPTV
>Retired in 2015

You have to keep moving forward. If he had refused to move forward from crossbar switches and 16mm projectors, he would have been left behind 25 years ago.
I do believe that we will eventually reach a point of critical unemployment, but that's another discussion.

>>62142777
>wageslaved for 25 years and retired to luxurious lower middle class existence from his meager savings
living the dream

>>62142909
you don't need much for happiness

IT wages are still 2-3 times higher than the average and IT people tend to be better investors/smarter spenders which helps too

>>62142909
What's your alternative?

>>62143809
Move up the ladder, network, go into C-suite or start a business, make actually good money, build something of your own, invent or research something new. You know, all those things sane smart people do instead of working paycheck to paycheck as a glorified technician sucking your superior's cock.

File: 14821.png (21KB, 200x194px) Image search: [Google]

21KB, 200x194px

>>62140597
This will change soon. Albeit, the current methods will be BTFO which is why it wasn't smart to have chased the statistical optimization meme degrees and approaches

File: dragon_arising.png (1MB, 1119x839px) Image search: [Google]

1MB, 1119x839px

>>62140668
Soon. Also, the correct software can drive hardware anyway it likes.

File: 1501043625322.jpg (76KB, 701x870px) Image search: [Google]

76KB, 701x870px

>>62140752
Decades? Think on a much shorter timeline...
And yes, this is incompatible with current approaches

>>62143938
Please don't tell me you bought the Startup meme. Good luck sucking VC-Fund cock while you preach to the choir you're a "self made man". The sucking a cock mindset is bullshit, you can make good money doing contactor work, the whole salvation by owning a company meme is bullshit. Even if you sell it to Google you only make pennies

>>62143978
>starting your own business is a meme
>literal I'm 14 and read TechCrunch once autism
Please, stop posting.

File: 1491255558070.png (298KB, 515x438px) Image search: [Google]

298KB, 515x438px

>>62140833
ML is set to get BTFO

>>62140909
Quantum computing is another misapplied tech that's being directed towards statistical bullshit models. Don't expect much from it until someone gives the right blueprints and models.

>>62141614
^this

>>62141632
Yes, this is the industry norm and its why they've gotten nowhere in their ivory tower ventures but local minimums.

>>62141669
Rarely do you develop algorithms. There are papers that are 40 years old that they're still dusting off. Instead, you babysit and tweak models like a code monkey brainlet. What's not disclosed outside certain circles is that they hire PhDs mainly to pad the payroll and demand higher valuations and capital. If it were know that undergrads could do the work, valuations would plummet.

>>62141680
>>62141717
^this

>>62141938
^and the reason why you need high qualifications to do any interesting work is because academics have purposely infected it with obfuscation and over complexity. There's a reason you never let a PhD write code or never see it published in their papers.

Good times

>>62144024
>>Anyone who is not working in "AirBnB for Cats" or some shit, sucks on his bosses cock and lives paycheck to paycheck.
Yeah, no

File: 6ee.jpg (40KB, 491x491px) Image search: [Google]

40KB, 491x491px

>>62143978
>yes, yes, doing things is yourself is gay and stupid and for plebs
>better work for (((us))) we pay you a great wage

Can I get a fintech job with an ML masters or do I neet a math/stats/probability masters and nothing else?

Deep learning is a meme, but I'm sort of hopeful that the hype may somewhere lead to a real breakthrough in AI research.

File: sad apu.png (37KB, 1127x685px) Image search: [Google]

37KB, 1127x685px

>>62141938
will MSc in Statistics get me good job

pls

>>62144081
>>Of course goyim, have all this money for your company, remember I know own 60% of your soul. My doorman is a such a strong blackman, why don't you marry your daughter to him. Multiculturalism is the future! Like your startup!

>>62144060
>the only kind of business one can start is making memeapps and hip webservices
Am I being trolled or are you legitimately that detached from real life?

>>62144115
Mah nigga, look at 80% of "trendy" startups. They're all fucking webapps. What are you planning to do? start a designer shoes line?

>>62144115
In what business category do the 5 biggest companies by market capitalization work?

>>62144097
Fintech is everything from alpha testing in some shitty wallet startup and vomiting abap code to derivatives pricing and programming RT systems.

>>62144135
I'm not planning on becoming the next Colonel Sanders...

>>62144135
>starting your own business from scratch
>biggest companies by market cap
How is that even remotely related besides your autistic contrarianism?

>>62140123
It has proven a lot of useful real world applications just in mere experiments, there are imagined hundreds of real world applications just with the tech today done better.

>>62144152
I'm talking about quantitative analysis at a bank, hedge fund or algo shop

>>62144170
Just check the list. Half of those started up as a shitty web application. To dismiss the power of webapps nowadays is to dismiss the most powerful force currently on the market.

>>62144135
apple is the biggest and is at root a hardware company

>>62144192
>Apple
>Alphabet
>Microsoft
>Amazon
>BerkshireHathaway
>Exxon
>Johnson&Johnson
>Facebook
>JPMorganChase
>Wells Fargo

only two of those started as web applications

>>62144182
You can try, but chances are slim. Big boys only hire PhDs and smaller fish get a fuckton of PhDs anyways, because these places are also very lucrative for mathematicians and physicists and often they are a much better fit than a pure CS-ist.

>>62144224
Fuck. So I should try to get a math or stats MS then PhD if I have a CS bachelor's? I don't even know if I'll grad programs in math accept CS majors.

>>62144192
>desperately trying to steer conversation to another unrelated topic to continue being autistically contrarian about things you don't understand
Okay, this is my stop. Have a good one.

>>62144238
Depends on your uni, courses, credits, country, grades. Personally, I think you shouldn't do it. QA is not a developing with a lot of math job, it's pretty much near-research tier math with a bit of development job. Unless you're naturally passionate about it you're going to hate it. And while it's very well paid, it's not really a very cushy job with ib-level working hours and stress.

>>62144338
Are there any other career paths in CS that are as well or similarly well paid?
I heard quant firms also hire developers for their back end but they're probably not paid nearly as well.

>>62144376
Enterprise, RTC, critical embedded. The first might make you want to kill yourself from boredom, the latter from stress. Choosing just for salary will not bring you happiness. Try different things and do something you like at least a little bit.

>>62144429
>critical embedded
This sounds interesting though, what are some concrete examples? Aerospace firmware?
>Choosing just for salary will not bring you happiness
I just want to save up money and retire as early as possible.

>>62144482
>concrete examples
Any and all critical systems really.
>retire as early as possible
Well, now I feel bad for giving advice to an autistic moron. Good luck anyways.

>>62144523
>he wants to keep working until he's 70
Good luck to you too.

>>62140073
Are you interested in his new course specialization in coursera?

>>62144604
No