Thread 106250710

worth it /g/?

depends on your usecase, which is virtually none.

>>106250831
how is getting a job for a usecase

I have a lot of opinions about this and I can send a long-winded vocaroo message. But in one sentence: no, don't get a Data Science masters, and just get a CS master's instead and focus on data engineering pipelines that feed into a data science analytics layer

>>106251030
Come from a nontechnical background. Probably couldn't swing an entry into any halfway decent CS master's program.

>>106251128
Dude, you are talking to someone who went to a shitty state school and got a bachelors in psychology, then pivoted fully to data science and then eventually full stack dev / normal SWE work.

Just go directly to CS, skip the DS. Data engineering is also way more valuable than data science for enterprises

>>106251169
Have you completed the degree? What were your job prospects following

>>106250710 (OP)
>Data Science
No.
You'll get smothered by an army of pajeets
> all lie on their resume
> none of them have any real, tangible experience
> it's a shitty field because there will always be somebody that will work cheaper
> "but my work is top notch"
> think about it, you're working for bean counters
> ALL they care about is cheapest
t. was in the data science field and jumped, it's awful

>>106250959
Not in Trump's (Satan's) economy. Lmao

>>106251030
This is good advice. There are very few Data Scientist jobs, and meanwhile there are around a billion Indians who used scikitlearn once and now consider themselves to be a data scientist.

t. Came from a previous non-tech career, got a masters in data analytics, and now work as a data engineer. It's comfy.

>>106251307
I did not get a formal CS or DS degree
>got bachelors in psychology
>did data science bootcamp
>got data science job
>at this data science job i realize how useless data scientists are to companies compared to normal software devs (full stack) or data engineers for companies with large data warehouses/pipelines
>did a shit load of side work related to full stack dev
>pivoted to SWE

No degree involved.

BTW, I literally am the Indian pajeet who did data science lol. I'm Indian-American tho. But holy fuck, it literally is an army of pajeets in data science just being ape retards with a jupyter notebook with the scikit learn SDK And plotly.

I want to leave a vocaroo about this so I will, here's a more long-winded explanation for why data engineers and/or full stack devs are way more valuable to *most* companies than data scientists:

https://voca.ro/1aohRHP7dMDv

>>106251882
thanks for the response anon.

How easy was it to pivot with the data science background? I have a portfolio of projects and I'm a semi-competent programmer, the degree is mostly just trying to get a leg up in terms of finding a tech job. I don't foresee data scientist as a future career for myself I would rather leverage the degree to branch into other technical fields... possible swe

>>106250710 (OP)
As an international student in my last semester of Software Engineering (and no I am not a jeet but god damn there's a lot of them here even the professors), I’d say a master’s is only worth it if you enjoy research. If you think you’ll learn programming from a master’s, not really, unless your thesis focuses heavily on it, and even then it’s still academic formality. If you want to program, focus on your language, get certified, and build real products and while you're at it, study Project Management (agile, scrum, etc.).

>>106251882
>https://voca.ro/1aohRHP7dMDv
you don't sound indian, but I agree

The data fields also seemed like they had been dying for a while since the mid 2010s, tools like scala, or spark popped off as data was the New Gold, and then people found out that these thing are expensive, and hard to interpret and that even if you get the best insights it all hinges on execs and managers actually making critically informed decisions. Finally, the hyped died down a little and big data was mostly left to the big players.
Of course, with the GPT boom it has picked up again big time for data engineers.

but yeah like he says, better to keep in mind the hierarchy of this. If all of this data turns out to be useless, or not worth the ROI of handling it. Then everybody gets axed except for the thing that keep the business running

>>106252181
doing data science with the goal to pivot to swe is very dumb. If you are going to do a master then do one in CS. If you want to do webdev then you probably don't even need a masters. Otherwise take different classes to see what you like. It's biased because it's what I like and clicks with me the most but I like computer systems subjects. Cyber could be a good one as long as you get computer system knowledge that goes with it (networking, OS, hardware attacks)

>>106251533
Digits checked
>army of pajeets
100% this
>be me
>data world, frenemies w/ PowerBI & Snowflake etc
>"anon, we need a certification program"
>ok
>multiple-guess quiz
>"anon, we want automated scoring, with a practical as part of the cert"
>ok, and ok
>biz dev is really happy w/ number of "partners" (offshore pajeet job shops) passing the cert
>I notice surprising clustering of scores among partners
>most students at a partner get the same score ... and miss the same questions
>theory.jpg
>I modify the lab
>I inject a couple of rows of derived data into each student's lab
>each student has a slightly different dataset to play with
>first couple of questions (number of widgets sold) are same answer
>lab gets more complex w/ normal dimensions (widgets sold by color) and time dimensions (widgets sold by quarter)
>by looking at test scores, I can tell the "source" of the cheating, the one that did the work and got it right
>and I can identify the cheaters with a high degree of confidence, the ones that blindly copied the source's work
>I let this cook for a few weeks
>present my findings
>one biz dev guy, "fuck! I knew it! they couldn't do dimensional modeling if their life depended on it"
>another, "that explains the constant stream of basic fuck-ups at $large_customer
Ended up leaving that place when the CEO took his shirt off on an all-hands call.

>>106252263
>Ended up leaving that place when the CEO took his shirt off on an all-hands call.
KEK, don't care if fake twas very funni

>>106252181
I'm just gonna fucking say what I would tell my child or close friend or other family who wants to get a fucking JERB in this field:

BUILD A REAL FUCKING PROJECT WITH A REAL WEBSITE AND A REAL INTERFACE WITH A REAL API OR APIS AND EXTERNAL API DEPENDENCIES

And most importantly. MOST FUCKING IMPORTANTLY. I am deliberately typing like an emotional female zoomer for a reason. Because the following is super important:


USE REAL AUTH APIS
YOUR PROJECT APP MUST HAVE A LOGIN / SIGNIN FLOW
USE REAL PAYMENT APIS (you should prolly use stripe API or something)

USE REAL FUCKING CLOUD RESOURCES. Learn how to SSH into a VM.

LEARN DNS

LEARN ABOUT THE CLUSTER FUCK THAT IS LOAD BALANCING

LEARN ABOUT HOW TO STAND UP A WHOLE FUCKING WEBSITE AND LANDING PAGE

LEARN ABOUT HOW TO CONNECT A DATABASE (e.g. AWS RDS) TO YOUR APP VIA ENVIRONMENT VARIABLES

LEARN CI/CD - LEARN HOW TO INTO THE WHOLE FEATURE BRANCH -> MASTER/MAIN -> STAGING -> PROD PIPELINE

DO THIS ALL YOURSELF DO NOT BE AN ACADEMIC PAJEET. You have to just to this shit yourself on LEGITIMATE home projects, not le epic streamlit app that data scientists do

I'm just describing what it takes to learn how to do all that shit it takes for a software engineer to build something that could make money.

>>106252363
Oh also learn QA - testing automation and how it ties into CI/CD.

Jesus fucking christ

>>106252363
>I'm just describing what it takes to learn how to do all that shit it takes for a software engineer to build something that could make money.
You just forgot two small steps:
>find the most useless meme idea of a product the doesn't need to exists (let alone exists as a webapp)
>hire marketers to shill the meme idea
Then you can start making money