How I became a Data Scientist at IBM (No CS Degree / No Masters)

what's happening guys in this video we're going to take a look at the exact steps that i used to become a data scientist at ibm so first things first my experience before i actually went into data science now i actually graduated from the university of technology sydney with a bachelor's degree in finance and accounting that being said there was absolutely zero technical components in that so no programming no coding and pretty lightweight maps now i leveraged that degree to be able to land a graduate program role at the reserve bank of australia this is the equivalent of the federal reserve bank in the us or the european central bank in the european union now the great thing about this is that i got to meet a lot of really intelligent and talented graduates as part of my cohort now just so happens that one of the other graduates that i was speaking to was actually taking andrew ing's machine learning course at the time and he mentioned something about supervised and unsupervised learning and i made a wisecrack joke which was kind of seems a little bit idiotic at the time but this actually got me thinking what is data science machine learning and deep learning so from that one interaction i began to do a ton of research into the field i wanted to take a look at the pros the cons the job market and potential in the future as well as what type of work data scientists machine learning engineers and deep learning engineers do and everything that i read sounded absolutely amazing so from that point onward i decided to make a decision that come hell or high water i was going to become a data scientist and every step that i took from here on out in my career was going to be in pursuit of landing that dream role now the organization that i was working out at the time placed a high value on advanced degrees there were a number of people within the organization that had received their masters and phds from universities like oxford mit and harvard so i figured maybe they know something that i don't and maybe i should go and apply for a master's degree myself so i went and did exactly that i managed to successfully get into a masters of data science at my alma mater university of technology sydney and enrolled in my first course and this is where all the problems began you see before i'd gone and applied for my masters i'd done a ton of research into the field itself i looked at popular languages and advancements that were happening in fields like deep learning natural language processing and computer vision in my mind i had it stuck that python was the language to learn this is where a lot of innovation was happening now there was a lot of conflict when i enrolled in that first course and the entire thing was taught in art i quickly questioned my tutors and asked why we weren't learning things like tensorflow or an ltk and why we weren't coding in python given a large majority of the advancements happening in the field were happening within those languages now admittedly that probably was the wrong approach to take but the outcome was probably the right one so i finished that first subject and managed to get a high distinction but i decided that university probably wasn't quite the route for me i decided i'd defer a single semester and see if i could save myself the sixty thousand dollars the course was going to cost and this is where my journey of self-study began now in spite of the fact that i deferred my master's degree i was still allowed on university grounds now this definitely played into my favor you see as part of that first course that i took in my master's i'd already established the habit of going to university on saturday and sunday and studying for a number of hours to make sure i was up to speed with the content now because i didn't have the course of study set by the masters anymore i had to come up with a replacement so i bought myself the book python crash course by eric mathis and set myself a goal of doing a pomodoros on saturday and eight pomodoros on sunday i tracked each one of these pomodoros using an app called promo to do the goal was to get proficient in python so that i could begin learning machine learning data science and deep learning this took roughly four weeks in order to get through the entire book now what i also did at the time was i built up a memory path this allowed me to learn all of the different concepts and retain them in my head as i was going through and learning each one of those concepts now if you want to see a deeper video on how i actually did this i've actually got a video on how i use specific memorization techniques to learn how to code i'll link to it up there once i finished python crash course by eric mathis the next book that i began learning from was hands on machine learning by aurelie bierhoff excuse my friend now i did the exact same thing for that book a pomodoro is on saturday and eight pomodoro's on sunday and i built a memory part that covered data science machine learning and deep learning but by this home i was itching to get started with some real projects and actually applied the skills that i'd learned by this point we were five months in the first course of university had taken three months and i'd spent two months learning python data science and machine learning by myself as part of my pomodoro process each weekend i told my manager about the new skills i was picking up on the weekend and asked if i could build some projects and automate some stuff for our team in my own time he was super open to this i began to automate reports and build visualizations using python i also built some time series forecasting models to help with our financial budgeting and forecasting now it just so happens that i got lucky at this time our team was implementing a new system for financial simulation and modeling for tm1 this would help our team with budgeting forecasting and a large number of our financial processes that we had to go through we brought in consultants to help with this implementation and i got the bright idea that maybe i could learn a thing or two from them so every night after work i would stay back and sit with one of the consultants and try to learn what they were doing as part of this process i learned an absolute ton about logic control systems security governance and data modeling in general i also managed to pick up a little bit of sql along the way i stuck around for around about six months while this project was being implemented and along the way i accumulated a ton of experience so much experience in fact i was able to leverage it to actually become a consultant implementing tm1 this opened up a whole new world of opportunity the consulting company that i was now working for didn't have a data science division so i figured if i couldn't get into data science here then i'd bring data science to me i began building machine learning models for our clients after hours to show them what was possible they didn't ask for it but i built it anyway because i knew that there was value there that they'd be interested in at this point the university of technology sydney was knocking on my door asking me how many more times i'd be deferred i deferred a few more times but i also enrolled in their startup accelerator to see if i could maybe build a data driven startup and out of this my startup was born now you're probably thinking nick you said you wanted to become a data scientist why are you building startups well the startup that we decided to build was all to do with leveraging emr data this stands for electronic medical record data the goal of the startup was to smooth the transition of patients throughout their journey in a hospital in order to improve their patient experience this was all data driven we used the emr data to better predict what their next step throughout the organization was likely to be ideally in hopes of reducing the number of readmissions and boosting the patient experience overall now unfortunately the startup completely failed and blew up in our faces but it was a great project that i was then able to place on my resume i've now spent a year and a half in that new consulting gig focused on building financial models data engineering pipelines and when i had the time machine learning models by night it also just wrapped up the starter and this is where i had my second bright idea i decided that i'd get involved in social media but specifically social media focus on data science so i built myself a wordpress blog and registered the domain name nicholasrinot.com the first blog post that i wrote was how to build an image classifier that would differentiate between batman and darth vader i'll never forget the look that my boss gave me when i told him that that's what i'd been working on but to be honest i didn't give a i was working and building data science model also at this time uts again came and knocked on my door and asked me if i'd be deferring again i figured that by this time it was probably the right thing to do and actually drop out so that's exactly what i did but to be honest i didn't really care i had the strategy that i was going to use to stand out i was going to write blog posts to show technical eminence and i was going to record youtube videos to show that i was able to communicate these concepts effectively if there's one piece of advice i can give you if you're gonna go down this track it's that you need to show what you're capable of this could be through twitter posts it could be through recording youtube videos it could be through blog posts could be through demo apps that you build whatever it is you need to show how capable you are this in the eyes of the recruiter is going to help you stand out head and shoulders above the pack so that's what i did but i took it one step further at this point i began networking i showed up to networking events and introduced myself to the players in the field now one company truly caught my eye i turned up to a data science event which was being hosted by ibm where they were showing their new product called data science experience now it just so happened that they had a data science there that was explaining the product it was showing how it was possible to use it to accelerate data science workloads so i approached him and i networked my ass off now there's a few key tips that you can use when you're going and networking at these events ideally if you know who's there and who you want to go and approach you want to do your research into find out who they are what they do and what's important to them have something that you want to ask them to truly extract their expertise but also you want to have your own elevator pitch prepared for you what you do and why should they care this establishes you as a reputable contact and someone that they'll be able to reference later on in the future so i did exactly that i explained that i was a consultant working in the industry i was building data science models and i was actively engaging in the data science community it just so happens that this contact would really serve me well in the future in fact he's one of my team members right now now this brings us to the painful bit in order to get my one interview i applied not once not twice not three times i went back and calculated every single application i made and in total i applied 16 times to get just that one interview i guess the key takeaway in this is that even if it takes 1 2 3 or even 30 applications keep on going you're going to get this every single application that i applied for i tailored my resume to focus on the technologies and techniques and specific industries that they will focus on as part of the consulting job that i'd already worked in i'd spanned a number of different industries i'd worked with a number of different clients but this really worked in my favor if you don't have experience in a field try and get some feel free to volunteer look for startups that are looking for junior developers or junior data scientists that is the easiest way to get expertise now when it comes to writing your resume ideally what you want to do is ensure that you keep it short ideally no longer than a page and every piece of job experience that you have on that resume highlight your achievements and what you did to truly stand out last but not least you want to ensure that you focus on demonstrating what you're capable of remember i said just a few moments before that you want to show your recruiter what you're capable and competent in this is where the youtube videos that i recorded and the blog posts that i wrote really came in handy and i'd highly recommend if you're going to go and apply for a competitive role like a data scientist a machine learning engineer a data engineer or a software engineer do exactly that because it is going to make you stand head and shoulders above the rest now once i got that interview i was in i was going to nail this i went and prepared like my life depended on it i prepared a cheat sheet that had just about every data science concept that i could possibly be quizzed on i went and researched potential interview questions there are a ton online and i'll link to some in the description below and i quiz myself every single night leading up to that interview i would quiz the absolute crap out of myself to ensure that i had appropriate answers for really complex data science concepts and for different things that i'd actually gone and built what i normally do as part of these interview processes is that i'll prepare five different case studies that i can reference whenever i get asked a specific question if you get asked about how you might approach a specific problem you want to be able to have a case study to back it up so on top of all the technical content that i had as part of my cheat sheet i also had my case studies laid out i also had a couple of different programming questions that i could use to demonstrate my expertise now on top of all of that i researched ibm as a company i took a look at what their main products were who their competitors were and what truly made them stand out as a company i also went and read their financial statements to see what their financial performance was like to ensure that i could talk about it if i was quizzed on it i was prepared i had a cheat sheet that was roughly about three or four pages long and everything that i could potentially get quizzed on was in there this brings us to the interview i had two interviews to get through in order to land the data science job at ibm so close guys the first interview was with the senior manager of the division that i was applying for which was data and ai he asked me a gamut of behavioral interview questions but also quiz me on things like what clients i'd work for what industries i'd worked in as well as what technologies i was familiar with now it just so happens at the time because of the startup that i'd built i'd also picked up javascript as a different language this allowed me to say that i could do data science projects but i could also build full stack apps this would allow me to show a range of different technologies out to their clients which really helped stand out the second interview was where it got really interesting i was interviewing with what would become my direct line manager but also the data scientist that i'd met at that previous ibm event what are the chances right it got a little bit trickier in that interview i was quizzed a bunch on algorithms data science techniques how it handles sparsity and unbalanced data but really these were traditional data science questions that you should be able to handle out there on the internet now luckily i was able to really shine in that second interview i'd prepare to turn it and i was ready to handle just about any question they threw out at me i successfully nailed that interview and within the next three days i heard back from the hiring manager which would now become my direct manager that i'd successfully landed the job now the cool thing is that within a couple of months of starting within the company i was actually flown out to singapore to take part in their pi r data science training so i actually had two weeks of dedicated data science training as well which was actually really really awesome i've got to meet some of their senior researchers as well as some of their data engineers out in the singapore technology park but that in a nutshell is how i landed the job as a data scientist at ibm so again it's probably not the traditional rat probably took me around about two years and two months or two years two and a half years it took a little bit of time but probably no more time that i would have invested in that degree anyway that being said it wasn't without hard work just because i didn't go down the traditional computer science or masters of data science degree path didn't mean i didn't have to put in the hours the effort and gain the experience in order to land the job that's the thing either way you're probably going to have to put in a fair bit of hard work in order to get one of these coveted positions whether it be a software engineer a deep learning engineer machine learning engineer or data scientist either way put in that work and i'm sure you'll get there and if there's any way that i can help along the way by all means do reach out that being said hopefully you enjoyed this journey and you got a little bit out of it thanks again for tuning in peace

Channel: Nicholas Renotte

Views: 1,625

Rating: 4.978261 out of 5

Keywords: data science, how i became a data scientist, data scientist, data science work, how to become a data scientist, how to get a data science job, data science career

Id: 2UXWNsOPN30

Length: 14min 48sec (888 seconds)

Published: Wed Oct 06 2021