A Data Scientist says, Don’t Become a Data Scientist!!?

hey guys welcome to the video very quick impromptu livestream yeah it's now 1:30 so I'm just gonna wait for a few people to get on board I'm not sure how many people are here yet software is not giving me an indication so we're gonna get into becoming a data scientist and comparing that to software engineering I found an article today or a few days ago rather on this subject and I decided to read the article and I thought it was an interesting piece so I thought we'd cover it here so I'm just gonna wait for it alive people to show up this is a bit of an impromptu you live stream if you're watching the replay just forward a little bit to get to the action I'm just want to wait for a few people to get on hey Steve how you doing I hope everything is well we only got 34 people I'm gonna wait till we get about 75 people which should be within a few seconds and then we'll be good yeah what's up what's up a duardo what's up Steve Dr 99 a take hey hey what's going on guys girls out everything is good can you keep up with both keep it up data scientist versus software engineer well we're gonna read this article and then we're gonna jump into it a Darian how you doing and then I'll do the Q&A so cover the article I'll make my comments as we go along and then I'll do the Q&A how would it how many people are out 50 people so we get up to about 75 and I jump into this subject I want to watch this but it's tiny little bit here goodnight but I am gonna watch this cuckoo hey guys hey Rob dee I'm not gonna say this right divya sharma my staff how you guys doing okay what's up what's up what's up all right so what do we got now we're at 62 what I hate about 7075 people which will be I would imagine very soon we'll jump into the subject at hand I'm good I'm good very busy now getting ready for my move so it's taking up a lot of my time that's my output is dropped but as soon as I move at the end of next week I imagine my output should shoot right up I have a whole bunch of ideas in terms of new content even some new courses coming out good morning from California hey Marco how are ya chu it would be around 10 o'clock no 11 o'clock for you yeah good morning what part of California from I have family in Altadena Pasadena alright he just started freelancing very cool very cool alright so what are we doing here we got 70 alright so I'm gonna jump into the subject and then we'll do some Q and hey alright guys so let's jump into it hold on hmm don't become a data scientist so I found this on this site here and I thought it'd be interesting read and we'll just take it from there so let me mute the popups okay so don't become a data scientist the advice I give when somebody asked me how to get into data science he says become a software engineer so this guy Chris you'll see why the second why I decided to to do this article here this is an opinion piece I'd love to hear your counter arguments below everyone and your grandmother wants to be a data scientist but Wow data science may be the sexiest job of 21st century I don't know about that but anyway that discounts another rewarding highly paid profession software engineer I often get messages from new grads and career changers asked me for advice on getting into data science I tell them to become a software engineer in stead so let's jump into us and see what he has to say as you know if you've been watching my videos for a while you know I did a video a while back where I suggest that you're better off probably getting into web development or development data science my main reasoning was that it would take you years of training to become a data scientist whereas what development you can jump right into it without a degree so you can start earning right away making a lot of money but we'll see what this guy has to say whoo happens to be a data scientist alright so his first point number one there are more software engineering jobs that's something I pointed out too I believe where's an order of magnitude more jobs and software engineering compared to data science below are a couple screenshots after googling data scientist and software engineer jobs so you're looking at data science jobs seven thousand six hundred and sixteen software engineer jobs fifty-three thousand eight hundred and ninety three so yeah you see the opportunity of course in terms of abundance is in the software engineer so let's continue at seventy six one six data science jobs compared to fifty she fails at fifty three thousand and change software engineering jobs this is just jobs in the US but other countries showed similar results supply and demand is a basic concept in economics and developers should always pay attention to that whether you're gonna work for yourself or work for some other people according to Glassdoor data scientists make more money but I talked about this another video that's kind of fake yeah I'll get into that later once I get into once I finish this article but my untested hypothesis is that dinosaurs JA diets excuse me but data science jobs are on are also on average more senior that's a very good point so data scientists probably have to go through a lot more training so they have a lot more to contend with so they have to go for a lot more you have to go through a lot more training so when they get into the job there are more experience so they at the start they may make more but meanwhile they have spent a lot more time school so let me just show this chart here according to Glassdoor data science make more money but I have an untested hypothesis so here we go so let me just continue down here if I said if you offered me a 1 million salary I'd open a I recommend you take it so apparently open a open AI pays a lot of money anyhow number two there's no consensus what data science means that's a very interesting point management often doesn't have a consensus of what data science means it's also possible by giving business constraints they don't have the luxury of rigidly following a framework of roles this means the responsibilities of a data scientists varies a lot from company to company see what a reasons I decided to do is article because you can clearly see that this guy here is now no longer an academic nerd he's actually somebody in the real world he started to see business considerations with regards to technology choices etc something I emphasized quite a bit in my YouTube videos my courses etc you have to always consider the market the business considerations whenever you're looking at any technology you're going to be jumping into something people don't talk about enough in my humble opinion so this guy I can tell right away but ok he's in the business world now so he's starting to recognize reality so so he continues this means responsibilities of a data scientist varies a lot from company to company so he's gonna lose chart a little quick sketch so it's interesting yeah so he's suggesting here that software engineer developer is actually solving real world problems where data scientists are more in the intellectual exploration role a lot of research that kind of thing detached from commerce perhaps shall we say that's based on his and based on talking to friends of mine that seems to be the case Wow while an ideal spectrum arose between software engineer to data scientist makes this it's unlikely that it follows in reality this particularly goes for scaling startups still building infrastructure hired Canada hired candidate end up working on problems a company currently needs solve rather on the role they may have been hired for so what he's basically saying is that you get hired for a job on a particular role but a lot of times you're gonna be just writing just code and just doing things so I had a dildo evidence let me just go back here scuse me and I'm total colleagues in the field is that many data scientists find themselves writing back-end code like software engineers I've known other data scientists who's crunched financials in Excel that's surprising this is this is in stark contrast to what you'd expect if you grew up on K go competitions I don't know what a KO competitions because probably one of those fake cold competitions and you know how I feel about cold Wars and all that kind of nonsense and when we continue data science is siloed yeah yeah most companies don't need as many data scientists as software engineers that's why you probably got a lot more jobs as for software engineers other companies are hiring their first out of scientists right now for this reason many data scientists end up working alone even if they sit at the same table as other developers this is important because it takes into consideration your work environment how do you want to work something you gotta consider don't just look at the money okay this can make it difficult to get feedback and second opinion software engineers either don't understand predictive modeling or are too busy working on completely different problems in contrast one of the perks of being a software engineer on a software engineer team rather is being able to say to colleagues I think we should implement ABC in X Y Z in the X Y in XY z-- that way what do you think be prepared to have a conversation with yourself or a rubber duck if you are a software engineer so it depends on your how you want to work you know so let's go to number four data science is exploratory is exploratory be prepared for awkward conversations with management regarding why something you spent two weeks on cannot be used again a friend of mine who's doing his master's degree in AI and so on he's it's a similar thing he's seeing the same thing working on solve versus unsolved problems when the fundamental difference between software development and AI let me go on bugs and constraints aside you know but you know whether most software engineering projects are possible before beginning any work that same cannot be said about ml where you don't know if the model will be effective - after you built it yeah you see a lot of a lot of dead ends and I know people aren't inside inside some companies are telling me the same thing I know a guy used to work for me his startup they did a big round of financing and they had a lot of investment in chatbots and AI and they had to scrap a lot of stuff after a couple years because they realized it wasn't going to work I have to maybe I'll cover that in detail in another video so let me get into it companies aren't ready for AI even in an era where every company is an AI company most don't have infrastructure support it it even or even needed the head of data science for rapidly scaled scaling startup recently shared some advice over coffee so when we read that again the head of data science for a rapidly scaling startup recently shared some advice over coffee first you figure out the problem then you build the infrastructure then you bring in the data scientist this is not a quick process another first data science hire at a well-known company recently rented to me she was forced to train AI models on big data on her laptop rather than in the cloud if that's slow if you're fraught in without specific problems to solve the company isn't prepared for data science you may find yourself struggling to add value yeah you know I seen before when you see the hype machine spin up for a particular new technology you see a lot of that and it's just normal it's a normal part of the process and you just gotta expect it as a developer you can make a lot of money in the process but you know a big part of being a developer is the rewarding part is putting out an app that works that people are using and you devolve it and you evolve the business that's what studio web is that's my sass my learning I architected it from scratch we built it out when it's we're on the fifth iteration version five I got lots of school clients and visual clients and we keep refining the software and the curriculum so that's a big part of fun getting that feedback and a lot of a IML is a lot of dead ends you hit but that's again that's normal I'm not criticising it's just a normal process when you're involved in new stuff let's jump back in software engineering teaches generic skills becoming a junior software software engineers like getting an MBA in technology you learn a little bit of everything you learn databases cloud deployment security writing clean code you learn to manage building software by watching your scrum leader senior developer or PM you get mentorship vehicle reviews you land in a company with an established engineering team it's almost guaranteed that you will level up your skills quickly and build a generalist background that's very good advice that's why I tell people you want to get out of doing tons of tutorials as quickly as possible and you want to get into the real world freelancer Z is low-hanging fruit just you know you know having colleagues is cool but just being able to interact with a client and having to solve a problem is a big part of the process of developing your skill as of developer no matter what type of development that you do and so I really highly recommend people jump into the game right away and as simple people have pointed out in commerce and YouTube and as I'll point out to you you learn much more quickly when you actually start doing real-world work that's why I structure everything accordingly in what I teach so let's jump back into this article hold on here we go software engineering is more transferable there we go a lot of flexibility is key key my friends flexibility by providing more holistic experience with technology software engineering provides better exit opportunities when you decided it's time for a change yeah basically when you be a software engineer there are a lot more jobs out there your have a broader set of skills perhaps and so there are a lot more jobs you could take DevOps security front end back at distribute assistant business intelligence data engineering data science I've known a number of developers who moved from software to data science if you skim data science jobs descriptions you immediately notice they're littered with core software development skills experience with SQL familiar with AWS Linux remember for experimental design for business experiments remember with DevOps such as get lab etc if you can build an end-to-end project you can also do more than build a model for Cagle you can take that model productionize it's set up as a ssin and stripe then start charging user access users for access rather that's your own startup i'd never argue about data science is isn't transferable making decisions based on data is a killer skill but it's also something that will become more of of every job as we come more data driven I would recognize that I could see I could see that where you would just have libraries you can tap into and they'll figure out how to do things and codify it and then you know where you go but to continue almost done here I think yeah Basheer machine learning will become a tool for software engineers number eight as AI commoditize --is is becomes commoditized rather and is easier to use software engineers will begin using it to solve their problems I can teach a developer to build school I know that is sklearn classifier in an afternoon I'm not a data scientist so alive my apologies anyway this doesn't mean they can this doesn't mean they can build the next alphago but it does give him an alternative to a hard-coded conditional logic based user input this is interesting that was one of the things when when all the AI machine learning stuff came about I had to sort of wrap my head around the idea of at the the the a I was actually figuring out what to do as opposed traditional most coding where you're actually just giving every step away you're telling the computer what to do as you know when you're writing software you're essentially telling the computer what to do except in the case of AI where you're teaching it you're teaching it how to do you're teaching it so can learn how to do it itself anyway dear scientists have specialized knowledge like statistics and intuition for how the model works but DevOps and security engineers have their own specialized knowledge as well yeah I just want to comment on this right here data scientists have specialized knowledge with statistics and an intuition for how models work that's something I was reading when I was reading about AI a couple years back and I was reading some articles from a highly experienced data scientist and they're talking about the machine learning how so like how they select the data parsed data if you will it's very important in terms of getting the whole thing to work and they talked about as being kind of an art form an art form being able to be very selective in terms of what to present to to the AI which I found kind of interesting all right what's going on let's continue I'd argue that these more common i scuse me I'd argue that these are more common than different a seasoned software professional could move between specialties an order of magnitude faster than a new entrant could pick up one again that's something a common theme I talk about all the time when people are always nervous about who should I learn this should I learn that oh boy if I learn node or should i do denno or no should I do not Ruby etc okay it doesn't really matter once your software developer or once you know Co won't you understand your fundamentals you can pivot all over the place so I wouldn't get to get too concerned about that it's a false fear you know why I don't think we'll see a complete merger wow I don't see excuse me I'm going to try that one last time Wow I don't see I don't think we'll see a complete merger of data science into software engineering it does feel like data science could become another software engineering speciality I agree 100 what I agree 100% with that all right number one AI is not replacing software engineers that answers a big question people are afraid is gonna come around and replace coders well think about it this way ai still can't drive a car down the street successfully right it can do it on the highway very limited you know but it couldn't drive down a complicated city street - you know we're a long way away you know years after an AI could drive a car successfully down the street with no problems and we're far from that Elon Musk's will even save us will tell you that long before AI even comes close to replacing coders you're gonna see cars that are totally and not autonomous level 5 autonomy right now they're at level 3 they go 4 and 5 is like orders of magnitude more difficult apparently as silly as it sounds I got into software engineering 2014 he's just a baby because I worried a I would render every job obsolete yeah yeah he got yeah yeah the AI hype you know came her out and came out to people right well no I it's gonna place everything it's gonna be a slow process all right yes since then but dial barely moves outside specific environments yep technology adoption is slow and AI is narrow narrower than the media would have you believe compared to other professions machine learning is even further from automating software engineering while we have startups building cool products like AI enabled code completion I actually had a sponsor right I forget their name also but they have an AI enabled co completion one of the smaller sponsors from a couple months back pretty cool tool writing code isn't a real job the job is solving problems using tech nology pre singularity that will remain a valuable and highly paid skills if you don't know what the singularity is that's when true AI comes around you know you have like Terminator AI comes around and instantly I mean comes like faster smarter than all humanity put together then we're in big trouble or maybe not but so there you go that's why let me just read this conclusion and then we'll go on then we'll do some Q&A excuse me conclusion first this is anecdotal secondly I realize I completed a scientist ml engineers research yes he did but I think these arguments are still worth considering given this is your career don't take it too seriously I prefer you research and make your own decisions that's part of being a data scientist after all at the end of the day were paid to solve problems so this guy his name is Chris I'll put the link to the article Centaurs data science comm I came across this and I thought it was a cool piece simply because it reflects what I've been teaching you guys for quite a while now right and I'm not a data scientist but you know I've been coding for three hundred years so I see the same patterns play out over and over again new tech if all comes out everybody hyped it up people flood into it collapses and then it kind of finds where it's gonna be the last big hype was the cryptocurrency hype right still out there I'm not saying it's dead but you know it's not the pass iam people were making it out to be initially right so I hope that makes sense all right so we're 23 minutes in what would took the time to write that article so let's see let's see if we got some questions here and I will get into it when we scroll back up where we go in here all right what's goin on what's goin on let's go home everybody as well hi I just buy your Python three cores Albuquerque thank you very much appreciate it you're gonna learn Python super quick now where's your hair where's your hair that's it I don't know one day I woke up but it was gone it was just that's it that's how goals sometimes it's how it goes 10:30 here in California I'm I'm in oside osai California I don't know what that is that Southern Southern California anyway yeah I I used to have super long hair and then I became a bouncer and then I got into a bit of a thing and somebody was like somebody came up and was grabbing my hair and pulled me anyway so the next day I cut off my hair and it's just been getting progressively shorter and shorter over the years the way it goes Valerie hi from Ukraine a that's where my family's from Kiev why did I come here that's a major I'm picking at USC this fall with this as I said you can't go wrong whatever specializations you get into you do so you you know you do your data science you come out your data scientists and you do pivot wherever you want to go you know I started I started an AI course in Python from Harvard very cool how's your quarantine time it's pretty good it's pretty good I because I run the business I own the business I always kind of work off hours sale for example once I'm done here it's about 2 o'clock my time I'm gonna go get a coffee you know usually 2:00 in the afternoon there's nobody around so I go and I go driving coffee I always wear gloves you know way before they opened a pandemic for last 10-15 years you wear gloves during the change of seasons if you live in a climate where seasonal changes chances of you getting sick diminish quite a bit like I hadn't gotten a cold or flu or anything like a decade since I started wearing gloves it's don't touch stuff you know let's see Kaiser says I've taken your foundation course what are some other good resources to use as a beginner I have a strong understanding the basics and can make super project but I want to do more what I would suggest that you do Kaiser is jump into the game it's time to get into the ring trying to get into the ring go out there and do a job for a small business even if you're not sure how to do everything you got your foundations when you got that in front of you you got a little job you got to do and even do it for do it for free a little job one week or two week job it's amazing it's amazingly instructional better than doing tutorials that's the plan hello Danielle oh how you doing all right that's one place I want to visit ADT however I hope the move is shooting me well yeah what's more I go in with the contractor to do some final little little minor rentals on the new place the big move is actually next week but I've been slowly getting things ready packing and stuff why doesn't a I finish the course for me exactly when that happens then that's it then where nobody's got jobs hey please what is the difference between software engineer and programmer in in practicality you know it's in the real world when you actually get our job it's whatever you know it's pretty much the same thing I find this software it's a different set in schools there's different set of courses engineers are more low-level theoretical level I'd have to do a total video on that but in in real world and that's all really matters at the end of the day pretty much doing the same thing you know if there are going to be too many data scientists out there demand will drop and salaries will go down stay in fields worker always demand and you can get good salaries yeah well you know again if you know how to code you just pivot to whether the action is the I goal got it does I think I actually pronounce that properly just watching from Brazil cool what part of Brazil are you watching from I like this look that's cool there we go what are we doing now right 28 minutes I'm gonna be ending this off and the three handle 30 minute I'm trying to keep these down to under 40 minutes you free me from all these Python hype and false hopes and guide me through right way that's why I'm here to see people have asked why don't you do tutorial no want to do a tutorial on this framework why don't you do a 1000 tutorial one because there's a lot of that out there a lot of nerdlings will teach you these things so what I try to bring to the table is my hundred and sixty nine years of experience in the software in the business world where I've actually made make money selling my software or I made money being a freelancer so I'm trying to teach you that because I think that's what's missing in the YouTube arena at least there's a much much less so it's plenty of tutorials to do but I try to bring the real world experience that's my goal here mmm do scientists use Python quite a bit they do use it quite a bit yeah if you're gonna choose like R which is languages use a lot in academic words or Python I will go Python again flexibility you want to be flexible so you can go in all kinds of different directions depending on where the money is where your interests lie etc python is very much used in data science yes exactly what else we got here DevOps or data analyst depends on what you know take a look see what you like to do always consider don't just look at the dollars guys because I've shown in previous videos that the difference in the salaries between the different technology coder types coders what a data scientist devops person they start to even out over time especially with a couple years experience so you know choose what you like to do you know really pay attention at because even if you're making extra five grand a year tenant year you know 100 or 150 thousand is nothing and if you don't like what you're doing and lice you good ah Python is the leading language in this field it indeed it is as far as I know by far I don't think it's going anywhere I don't think it's going anywhere yeah that is true that issue a lot more opportunity data scientist is so specific Vegas or maybe but software engineers very broad exactly oh here we go CAG okay goes free data sets websites they hold campaigns on data sets often there we go so there you go data scraping transforming and all I need is e store selling my services okay how are we doing for time all right we're hitting 30 minutes so I want to answer a couple more questions and then I'm gonna go get myself a nice coffee I might do a Instagram live but no guarantees all right what do we got here mmm this guy this guy des told Dayton is on the ball this guy he's going places for sure alright let's see what else we got here I'm always talking about here as a data scientist I can definitely vote for K go I think it is a legit learning platform there you go cool so you want to get into data science check that out I don't disagree with the statement data science is definitely here to stay and evolve but companies will keep positions to a minimum sadly they believe it can be a one-man team yeah I can see that it could be so it's you're confirming what that guy said right ok what else we got here see well see what Ken Kenji has to say I think the general data science role will start to disappear you will see companies move towards more specialized roles like data engineer data analyst visualization specialist and m/l engineers there you go and I believe Ken G's is into holy Sh it so cool so cool also you go ok well end with this week in week lat sing I know it smells ESS Jess on jQuery bootstrap PHP and MySQL and a couple of things I learned about it it's time to look for a job or do some more courses job job job job job job job job job cuz you're gonna learn learn much more doing a job - one or two free contracts so you get some reputation learn some real world skills and then you'll open up all kinds of doors in terms of jobs so yet nice time now signed to job it yeah yeah if your data set is wrong your answer is wrong yeah that's interesting my friends who who invested a bunch of money into that that's what they found out they found that the data was more it was super important the most important part of the in their opinion was the most important aspect of the whole ml thing again I'm not I don't know this the field very well at all so just superficial talking to people about it but that's what they say the data is huge why Stefan talking like his energy is sucked out because it is it is I need my coffee I only have one coffee today I'm slowly falling asleep is learning Shopify including theming and drop shipping worth the effort or does it require a lot of soliciting for clients and is the market flooded Troy depends on where you are what I suggest you do you basics then you go out to do freelance the reason I emphasize you do your fundamentals your basics so this way for you to drop to jump into Shopify or theming or whatever you want to do it will be much much much much much easier for you so you'll be able to pivot according to the demand you guys stop stop thinking I am a Shopify implementer or WordPress person or a PHP person our Python person you just think I'm a developer and as developer you can go and you know go in any direction that you want right depending on the demand take in that way then you'll be okay in my own career and I walked that talk I've written commercial code or been paid to code using eight or nine I've launched try eight or nine languages over the years and I won't even talk about technologies frameworks and libraries all kinds of weird stuff and sometimes as a freelancer I'd walk in and I would I would learn stuff that I had never used before I was just I was just aware that's why I've a series where I'm gonna be doing more inner news or gonna be doing some I'll be looking at different technology stacks and stuffs just so that you become aware so each aware of the possibilities when I first I learned just philosophy by the way from Bruce Lee Ji kundo which is you could argue is kind of the first MMA type of guy and his idea was to absorb what as useful as he said and every reject what's not useful for you and they'd come up with your own stuff and I kind of adopted that philosophy when it comes to software development coding you you want to be aware of what's out there so that you know what's out there like we learned you know different types of submissions that could be done to you and the way people might attack you or different tactics you as a fighter will use a very limited set of tactics probably and you may have one or two or three techniques you like to execute but you have to be aware of what's out there so they can't surprise you with anything in software development you want to be aware of all the you want to be aware of what's out there was available that doesn't mean know it but just be aware of what it is what it can be with it can do where it might help you solve a problem so when you run across a problem you're gonna go oh I can do that with Python and this module or oh this is better done with PHP and Arabelle all this is something I should never do with Ruby all right guys I ended up with the required we joke I hope you enjoyed this dream and I'm gonna let you go because it's time for coffee

Channel: Stefan Mischook

Views: 32,167

Rating: 4.8497653 out of 5

Keywords: datascientist, developerjobs

Id: -hk6nYLTTWk

Length: 37min 35sec (2255 seconds)

Published: Thu May 21 2020