Kirill Eremenko: This is episode number 261 with Data Science Writer, Andrei Lyskov.
Kirill Eremenko: Welcome to the SuperDataScience podcast. My name is Kirill Eremenko, DataScience Coach and Lifestyle Entrepreneur, and each week we bring you inspiring people and ideas to help you build your successful career in Data Science. Thanks for being here today and now let’s make the complex simple.
Kirill Eremenko: This episode is brought to you by our very own data science conference, DataScienceGO 2019. There are plenty of data science conferences out there. DataScienceGO is not your ordinary data science event. This is a conference dedicated to career advancement. We have three days of immersive talks, panels and training sessions designed to teach, inspire, and guide you. There are three separate career tracks involved, so whether you’re a beginner, a practitioner or a manager you can find a career track for you and select the right talks to advance your career.
Kirill Eremenko: We’re expecting 40 speakers, that four, zero, 40 speakers to join us for DataScienceGO 2019. And just to give you a taste of what to expect, here are some of the speakers that we had in the previous years: Creator of Makeover Monday Andy Kriebel, AI Thought Leader Ben Taylor, Data Science Influencer Randy Lao, Data Science Mentor Kristen Kehrer, Founder of Visual Cinnamon Nadieh Bremer, Technology Futurist Pablos Holman, and many, many more.
Kirill Eremenko: This year we will have over 800 attendees from beginners to data scientists to managers and leaders. So there will be plenty of networking opportunities with our attendees and speakers, and you don’t want to miss out on that. That’s the best way to grow your data science network and grow your career. And as a bonus there will be a track for executives. So if you’re an executive listening to this, check this out. Last year at DataScienceGO X, which is our special track for executives, we had key business decision makers from Ellie Mae, Levi Strauss, Dell, Red Bull, and more.
Kirill Eremenko: So whether you’re a beginner, practitioner, manager or executive, DataScienceGO is for you. DataScienceGO is happening on the 27th, 28th, 29th of September 2019 in San Diego. Don’t miss out. You can get your tickets at www.datasciencego.com. I would personally love to see you there, network with you and help inspire your career or progress your business into the space of data science. Once again, the website is www.datasciencego.com, and I’ll see you there.
Kirill Eremenko: Welcome back to the SuperDataScience podcast, ladies and gentlemen, super excited to have you back here on the show. And today I literally just got off the phone with Andrei Lyskov who is a Data Science Writer and a recent graduate who just got a job in Data Science. We had a fantastic chat and if you are looking for a job in Data Science or you are already interviewing or about to go into an interview in Data Science, you will find this podcast super valuable because Andrei shares not only his experience but also his research and his thoughts and ideas in the space of getting a job in DataScience and it’s really valuable coming from somebody who’s just recently been through this process himself. You’ll find plenty of golden nuggets which will help you get along the way. So here are a couple of examples.
Kirill Eremenko: We talked about the trichotomy of control, the importance of referrals, what kind of website Andrei has created for his profile for himself, and why. The importance of a portfolio and how Andrei goes about creating his, how Andrei taught himself Data Science, learning how to learn or master learning and lots and lots of other insights in this space. So if you are looking for a job in Data Science or you’re actively interviewing, this is the podcast for you. Plus Andrei will share a lot of additional resources which you can go and read so this, will be a great overview of those materials, and then in the show notes you can go and actually dig deeper into them. So with that said, can’t wait for you to be part of this conversation and without further ado, I bring to Andrei Lyskov, DataScience Writer.
Kirill Eremenko: Welcome back to the SuperDataScience podcast, ladies and gentlemen, super excited to have you back here on the show. We have a special guest Andrei Lyskov. Andrei, how are you going today?
Andrei Lyskov: I’m doing well and very excited to be on this podcast and to share whatever knowledge I have.
Kirill Eremenko: That’s super exciting. And so Andrei is… Andrei, you are a writer and you write towards Data Science. So you’ve got some very interesting articles. You’ve been featured by Forbes, Huffington post and others and recently, very excited to learn that you just got a job as a Data Scientist at Apple, where you’re starting from next week and it’s something that we’d love to dig into further as well. And yeah, man, thanks for coming on the show. How’s… you’re in Canada right now, right? What’s it called? A few hours east of Toronto. How’s things going over there?
Andrei Lyskov: Yeah, it’s great. I am now an official graduate of the university, which is nice. I got my diploma, so I’ve been hanging out at my parents’ cottage just enjoying the water and the nature before I move to the Mecca of tech in San Francisco as you had mentioned I’ll be starting at Apple. So it’s nice in between before I start getting really into coding and the whole tech world. I have lived in San Francisco in the past, so it won’t be completely foreign move for me, but the South Bay where I’ll be working in Cupertino will be a new home. So excited for that.
Kirill Eremenko: Wow! I did actually realize that you just graduated, when did you graduate?
Andrei Lyskov: I graduated about a month ago. The path to graduation had some twists and turns though I had actually taken a year off to work in Los Angeles at a company called Soylent, which makes meals in a bottle among other things, very popular with the hackers and developers as you can imagine. So actually, I graduated much later than some of my other friends.
Kirill Eremenko: Okay, gotcha. So, but even apart from Soylent, I’m looking at your LinkedIn and it’s got so many different roles that you’ve been in, like you’re a founding member at the Kairos Society, you were growth intern at IBM, you were a Data Science intern at Peer Street, that’s pretty impressive for somebody who’s just graduated to already have so much work experience under your belt. Like you must have been quite busy during your uni, you’re studying and working at the same time. How was that for you?
Andrei Lyskov: Yeah, I went into university with the mindset that school is a credential for me. As bad as that sounds, I had no interest in grad school. I had no interest in getting straight A’s or anything like that. But I was really interested in getting as many interesting experiences as possible and I found that university was great playground for experimentation. And as you mentioned that, I was able to basically expose myself to all kinds of different roles and experiences. Had summer where I worked in Beijing, I had a summer where I worked in Los Angeles. And these experiences were really instrumental for me because it made me realize that the confines of a classroom can only teach you so much. The real world often is a much better teacher. I optimized for those experiences rather than a 4.0 GPA or even getting involved in university life. I stopped joining clubs or starting clubs after my second year because I realized the payoff was not as much as getting involved outside of the university.
Kirill Eremenko: Interesting. Very cool approach. Very probably different to what many people take, and for me, I definitely thought about university. Okay, I got to get straight A’s and I was thinking, you might go… because I studied physics, maths, I might go into research further off and that didn’t happen. I moved on to other things. But how did you develop that conscious mindset and approach at such a young age? Going into university with that intention that I want to get the experiences over the straight eight grade or over the partying. Some people go into university to party, that’s another option.
Andrei Lyskov: Yes. I definitely indulge a bit in my first year with that, but afterwards I was much more driven or ambitious outside of that domain. And to answer the former question, I think a large part of it was the bookworm in me I suppose. I had read a lot of books when I was in high school, and the type of friends I surround myself were also like very driven and ambitious and we loved to practice self development. We would have weekly check ins with, you’d set these goals like, I want to go to the gym three times a week or I want to be eating clean for whole week or some variation of self development goals. And that really helped me see the incremental steps one takes each day towards a goal.
Andrei Lyskov: When I got to university, I had that mindset already where I was like, “Okay, I’ve seen this formula in the past where if I have this goal, I can break it down into daily steps and we’ll work towards it slowly in”. It might not happen next week, but over time I can start tripping away and working towards it.
Kirill Eremenko: Okay. Yeah. Very cool. You have an article actually that I looked through in on LinkedIn. It’s called One Simple Tactic From the Stoics For Setting Better Career Goals , and you talk about trichotomy of control. Can you tell us a bit about that, because it sounds quite in line with what you were just describing about how you said yourself, your goals, and for the purposes of our listeners, I think this can be very powerful for setting goals for your career. Whether you are still yet to graduate or you’ve already graduated a long time ago, and you’re building your career. I think that is a new thing for me. I hadn’t heard about it. This trichotomy of control philosophy, so yeah, please do share.
Andrei Lyskov: Yeah, absolutely. So the Stoics are these [inaudible 00:12:00] philosophers from back in the Roman times. They have a bunch of really famous philosophers that some people may have heard of like Seneca as well as this Roman emperor called Marcus Aurelius is who wears this philosopher keen. And he wrote this book called Meditations. For one reason or another stoicism has become really invoked in the tech world. You have everyone from Jack Dorsey to someone like Tim Ferris reading and promoting the virtues of stoicism. So, that’s a quick context on that. The actual tactic that they have employed in the past or I suppose mental model is called the trichotomy of control. And the idea is that in life there’s kind of three categories. You have things that you can control, which is basically you just take action on those things because you can control them.
Andrei Lyskov: Things that you have no control over, which is something you just let go of because you realize there’s nothing I can do about this. And then things that you have partial control over. And the partial control is where the crux of this article is focused on. So I’ll give you an example. You are preparing for a race, a 4K run, and you really want to win first place. So this is partially in your control. You can focus on your diet, you can focus on your sleep, you can focus on all of these levers. But at the end of the day you don’t fully have the ability to win. And so what the stoics recommend with things they have partial control over is setting internal goals. And so internal goals are basically goals that you have a much tighter control over. Whereas external goals, again, are less control.
Andrei Lyskov: So going back to that racing example, an external goal would be, I want to win first place. And an internal goal would be I want to be sleeping at 10:00 PM each day, getting 10 hours of rest. I want to be eating clean, drinking lots of water, sticking to my routine. And as you can see, you kind of break out this external goal first place into its sub components, right? How do you get to the first place? You have all these different things you can do to get to that point.
Andrei Lyskov: The idea is that when you do run the race and if you win it’s like, oh great, you know these internal goals led me to winning. Amazing. But if you lose you still are like, “Oh great, I still did my internal goals. So I’ve succeeded in that regard.” And so to kind of take this example and bring it to the Data Science domain and specifically with careers, you can think about this trope of working at a Fang company, which is like Facebook, Amazon, Netflix, Google, and a lot of people have this, I guess goal in mind, right?
Andrei Lyskov: They want work at this big tech company. They want to experience Silicone Valley and work with amazing perks and really smart colleagues. And that notion of an external goal can create a lot of suffering for someone because first off, these companies… To get in is, is such a luck of the draw. I think a lot of people don’t realize that… a lot of getting into one of these top companies is a matter of who was interviewing you and how they’re feeling that day. And so by setting yourself up for failure that way, by having this external goal or as an internal goal would be, “Okay, how do I get into a top tech company?” All right. You know, I need a good portfolio. I need a good resume. I need potentially find a referral, and obviously I need to have the technical skills so that I can pass the interview loop.
Andrei Lyskov: And so once you have those components in mind, you can break it out into daily goals where you might say, okay, each day I’m going to spend an hour on algorithms and data structures or statistics. And each week I’ll talk to three data scientists at one of these companies, so that I can start building these relationships so that when I do apply, I have someone that can vouch for me. And so even after you go through this whole process and you don’t get into this top tech company, you still will feel like you were able to accomplish your internal goals and you don’t have your hopes stashed or as the alternative is, you have your hope stashed and then you’re like, “Well shit, what do I do now? I didn’t reach my dreams. I feel like a failure.”
Kirill Eremenko: Very interesting, and even for somebody who’s already employed at a company where they love what they do, it could be I want to have a $10,000 raise by the end of the year. Right? That’s partially in your control, but not fully. Right? You don’t know what the [inaudible 00:16:53]… You might not… company policies might be different. You don’t know what your manager is thinking. You don’t know what the revenue of the company might be towards the end of year. But nevertheless, as you say, you can set yourself somewhere internal, what would be some internal goal examples in that case?
Andrei Lyskov: So in that case you want to identify past examples of folks that have gotten raises or promotions. Every company has a different way they reward their performers, some base it on, just how collaborative are you, how well do you get along with others. Another might be did you work on a project that delivered a lot of core value to the company? So as far as the first step in this example is figuring out which lever do you need to pull so that you can make a case for a raise.
Andrei Lyskov: The more practical route to getting a raise and this is funny enough, interviewing at other companies and getting an offer and then go into your current company and saying, “Hey, I have this offer competing from another company. Are you going to match it or am I going to have to leave?” And so there’s a lot of different ways you can get to this. And I think that’s part of why you don’t want to think about the internal goals, because there’s lots of different ways you can accomplish your external goal.
Kirill Eremenko: Very cool. Basically what you’re saying is have the… so there’s the what the why and the how, the Simon Sinek famous start with the why, I think there’s a Ted Talk presentation and there’s also Tony Robbins, start with the what’s and the why and the how. Basically have the what’s in mind, what do you want? Do you want a promotion? You want to start a company, you want to win a marathon. Understand the why you want that.
Kirill Eremenko: Really clearly is that really what you want and why. And then set yourself goals based… and then decide how you’re going to get there by. Set yourself what Tony Robbins for business pulled the massive action plan of how are you going to get to that destination from where you are now? And then, which you’re saying is don’t only reward or don’t aim to reward yourself or measure yourself, by that final goal that you were going to achieve the what, but rather reward yourself for every step along the way, in terms of how am I going to get there. Because that way…
Andrei Lyskov: The journey is only much better than this?
Kirill Eremenko: Yeah, exactly, the journey. It’s all about the journey.
Kirill Eremenko: All right, very cool. And speaking of interviews, you have another wonderful piece, which I’m congratulations as you, as you mentioned that when off the charts with thousands of views, the mastering the data signs interview. Since you’ve published it how many views has it had again?
Andrei Lyskov: It’s almost had 50,000 views, which when I wrote it, I honestly did not think it would go as viral as it did otherwise, I would have spent a lot more time editing and polishing it. But I think with writing, you just got to release it, otherwise it’s never going to get published. I’m glad it got out to the world and people are finding value out of it. I might do version two or some update to it just so that future readers can benefit from that. But I feel like folks that are just starting an interview can find a lot of value out of it because it does break down what you should expect at each step. And it also gives you some insights on how to best prepare for each stage.
Kirill Eremenko: Yeah. And so 50,000 views as 5 0 and you’ve just published it in January, so it hasn’t even been six months, a year and [inaudible 00:20:52] epic. Interesting piece and it basically walks you through the steps that one might encounter in a data science interview and I think it’s quite a good overview or acquire a useful overview you for anybody to keep in mind. And where did you get this information from or is this how the large companies do it? Is this… Did you poll people around? Like how did you compile all this together?
Andrei Lyskov: This was the synthesis of my own data science interview loop. I had started interviewing for a variety of companies at the beginning of my senior year. By something like November, I had done a bunch of different interviews, 20, 30, something like that. And each interview would proceed to various stages like the take home project, the technical interview, the onsite, and then obviously the negotiating stage. And so I was able to see each of these different stages. And before I even started interviewing, obviously I did my own research.
Andrei Lyskov: I talked to a lot of folks that have Data Science jobs and I pulled them and ask them, what is the Data Science interview loop look like in your company? How should I prepare for it? The reading online of course, the Internet has so much knowledge, it can often be overwhelming when you try to sit through everything. The article really was synthesized as a result of my own interview experience and there’s a way to consolidate this information while it’s was still fresh. But I also relied on other folks and their experience whether through personal connections or online.
Kirill Eremenko: Walk us through it. What are the steps somebody can expect in a Data Science interview?
Andrei Lyskov: Yeah, so you can imagine a Data Science interview funnel, this is the idea that, over your Data Science interview, month or few months, you’re looking at these companies, you’ll have this pyramid structure, that is upside down. And so as you go through this funnel, you’ll inevitably have dropout. Some people can just keep up a rectangle structure where they don’t fall out anywhere. Most people, obviously fall out at different stages, which is why the pyramid gets smaller as you go through each stage. And so the stages, I’ll give a quick overview. You have the coding challenge, which is typically unmoderated challenge so that asks you to do some pretty simple stuff.
Andrei Lyskov: It might be like a fizzbuzz question, it might be more complicated with a time series forecasting model. There’s a whole range of these coding challenges and so these are pretty easy to do obviously. They’re just making sure that whoever applies does not… actually knows what they’re doing. The second step is the HR screen, which is a lot more laid back then the subsequent steps. This is usually with a recruiter or some nontechnical person and they’re just making sure that you’re a normal person. As bad as that sounds, there’s a lot of people that get to this stage that drop out because they, they’re not able to communicate effectively or they’re unable to answer some questions, some really minor technical questions that are a recruiter might have. Like what is a windows function in SQL?
Andrei Lyskov: These are usually questions that if you don’t know them, it’s hard to get through it. That’s kind of the second step. Then you have the technical screen and the technical screen is really with engineering or Data Science folks. And this is the one where you might have an hour of conversations about data structures or algorithms or writing SQL queries. Every company has different way that they do these technical interviews. And so it’s important to do your research on various websites like Glassdoor or Blind, which might have some insight into how these interviews are conducted or what questions you should expect, even better is asking a recruiter, but they often don’t know what it is going to be asked, so they might not be able to tell you. But that’s the technical screen.
Andrei Lyskov: And then at some point you might even have something called a take home projects. And these have actually been pretty popular lately. They are used much more frequently in Data Science I think partly, because the science is not like software engineering where you can just ask a bunch of lead code style questions where it’s like, verse a tree or some variation of an algorithms or data structure question. Instead, they’ll give you a bunch of data, and they’ll say, do something interesting with this. Or they’ll say build a classifier. There’s a lot of these challenges that can take anywhere between 3 to 20 hours.
Andrei Lyskov: I personally love take home projects, I think they’re much better for me to show my skill, but I also recognize that I’m a new Grad that has a lot more time on their hands than someone that’s working. So, that’s a take home project stage. And then you have the onsite and the onsite, you will potentially present this take home project, or you’ll just be on an interview loop with your future manager or your future coworkers, might have your skip manager, which is your manager’s manager who will obviously ask more questions. And so at the onsite stage you want to be doing research on who’s going to be interviewing you because that will indicate to you what type of questions will be asked.
Andrei Lyskov: The nontechnical people, sorry, a data scientist with a social science background may ask very different questions from a PhD in computer science that has been doing deep learning research. There’s a whole gamut of questions that you’ll be potentially asked. So it’s good to do due diligence. And then obviously the last stage is negotiation and offer, which at this point, there’s a lot of different strategies one can take depending on your position. The article obviously gets way more in depth into what I just said, but that’s the quick TLDR.
Kirill Eremenko: Thanks. Fantastic. And we’ll definitely link to the article in the show notes. And this I think it’s very valuable again, and you [inaudible 00:28:04] yourself. I have been through interviews and well done for walking in that job that you are starting next week. Did this approach help you yourself and what did it help you do differently as opposed to what you wouldn’t have done had you not known about this? How this funnel breaks down?
Andrei Lyskov: Yeah, the funnel is obviously a general funnel. Every company does it differently. And so some companies might skip stages altogether. Like the Apple interview I did not have any coding challenges. I actually went straight to speaking to the hiring manager after which I went straight to the onsite. And then I also had a take home project during that process. Very different from my other interview with there was an insurance bank in New York that I was interviewing for and there was no take home. They had me to do the Skype interviews and then they flew me to New York and I had a two hour interview and that was it. And then I had very different experiences. But this mental model certainly helped me specifically the stuff about knowing your interviewers before you get to the onsite or you get to the screen.
Andrei Lyskov: That is something I did not do in the past. And it’s certainly bit me. The Internet, like I said, has so much information on it and that includes the people that are interviewing you. You’re able to do some light research to see who this person is, what school they went to, what their background is, what their duties are. You can really tailor your interview prep towards being your best self. Because Data Science is such a broad field, it’s really hard to know everything in depth and that’s why it’s important to do this due diligence. And I think that was really the biggest takeaway from the interview loop is I know who you’re being interviewed with when you’re going into this interview and what is expected.
Kirill Eremenko: Tell us about how much time you put into the preparation and how can others replicates our success and find these interviews like a breeze at the end of the day rather than a struggle?
Andrei Lyskov: Yeah. I’ll qualify that statement by saying that at the time I was taking four courses, one of which was the only course that I actually needed to attend and had any semblance of work in it. The other three were fallen 10 purposes, quite simple and didn’t require a lot of time for me. And so because of that, I had a ton of time to do prep and dive deep. And so for the take home project for example, I put in somewhere between 15 to 20 hours in the span of five or six days. Yeah, that was just the take home project. I also was doing some prep on SQL and statistics and python and machine learning. And so all included, I feel like it might’ve been 40 to 50 hours of prep just in that week before the onsite.
Andrei Lyskov: And that makes my situation a bit unique for obvious purposes. Not everyone has the luxury of that much time on their hands and whether it’s because they’re working or they’re in school with actually like difficult courses. In my case, I did walk in feeling like, “Okay, I’ve put it in a lot of time into this, I know this stuff inside and out.” And so I was able to really, walk out of the onsite feel like, “this went really great. I feel great. I feel like I answered all the questions exactly as I needed to and end up leading to job, which is also great.” But the internal goal is it’s just like spend the time, do my best, show up if that doesn’t work at least it was a good learning experience.
Kirill Eremenko: Okay. I love how we’re digging into this because I think especially for people who are looking to apply for jobs or a re about to have an interview processes, it will be very valuable. I’d like to dig into, to other components. You mentioned you had a referral and somebody who referred you to the company. Tell us a bit about that. Was it somebody that you knew a long time ago or did you make this connection recently and it just so happened or was it part of your search for a job that you knew through a referral, it’s easier to get you to a business and also any tips for people who are looking to jump on board a company without referrals. Because like from the stats out there I think it’s something astonishing.
Kirill Eremenko: Around 70% of hires actually happened through referrals, not through applications, but behind the scenes. Pretty crazy. Any tips you can share on that space? For us, this is your recent experience would be very appreciated.
Andrei Lyskov: Yeah, for sure. And to back up that just like every internship or job I’ve had has been through either referral, or I’ve gotten direct contact with the decision maker. I don’t think I’ve ever held a job or internship where I applied online and got an interview and got hired that way because you’re right, a lot of these jobs are being given to people that are being referred. And oftentimes, I don’t know if a lot of people know this, but when you apply to certain jobs, they might already be taken. And the only reason this job is on there is because there’s some rule that one must post a job before they make the hire. Or they interview people just to say they interviewed people when they already had someone in mind the whole time.
Andrei Lyskov: When you say it out loud it’s like, “Okay, why is this happening?”. There’s a whole bunch of stuff going on behind the scenes that are not going to get into, but that is the reality of the job market. So to answer the question of how does one go about getting referrals and being in a stronger position when applying to companies. So in my case, this was an individual that actually got introduced to me through one friend and at the time I was living in Los Angeles and this was a pretty good friend of mine. He had worked at Netflix and he knew someone that was a data scientist at Netflix. And they put me in touch with him because I basically decided like, “Hey, I’m looking to talk to data scientists. I want to make this transition to this career.” I don’t really know what I’m doing to give context at the time I was a business analyst at Soylent. So very heavy on the business end, not so much on a technical.
Andrei Lyskov: And so I was a fish out of water so to speak. And so I suppose the first step is you already have friends. Well, I hope you have friends or family or people that you can ask them and say like, “Hey, I’m interested in this field of Data Science. Do you know anyone that does anything remotely related to this?” And just get those warm introductions to these potential people. Because in my case, my friend definitely went above and beyond and vouching for me. And the introduction would have been a lot tougher if I didn’t have this person doing such a warm introduction. After that point, he and I would have regular check-ins. I was applying to Data Science internships and I got some feedback from him on my resume and then I started the internship and I had this problem with one of these stakeholders. I went to my mentor, more informal mentor of mine and said, “Hey, I’m having this issue. And over time I developed this relationship and I was never thinking this is going to lead to a job. It was always more like, “Oh I really want to learn” and this is someone that has a lot of experience. At some point it just came up that, “Hey, you’re graduating soon, if you need a referral, my team is hiring.” And that’s how it came about. And this is also how I got interviews at a couple of other companies as well. Well, there were a couple of times where I did ask for direct referrals, but the indirect referrals are often a much better method.
Andrei Lyskov: I think this long term strategies is much likely to yield fruit versus a short term thing that a lot of people have. Or it’s like, all right, I need a job, I’m going to connect with a bunch of people on LinkedIn, I’m going to ask them for a referral. That’s very hard to yield fruit from that. But a long term strategy definitely requires a lot more time and thankfully the person that I connected with it was like very nice, genuine person and it was very helpful.
Kirill Eremenko: And just to clarify, they already had the connection within the company that they were working to?
Andrei Lyskov: Yes. That’s right. He was already growing. They were looking to hire, he knew that I seem like a competent enough person where I wouldn’t embarrass him in the interview at least. And that’s part of it. It’s like you got to demonstrate that you’re not going to embarrass the person if they give you a referral. And so that’s an important thing to keep in mind as you’re thinking about how can I get a referral?
Kirill Eremenko: Great Advice. It’s the hard truth that it’s not a quick thing. It’s not a quick game. And the intention should be different. It shouldn’t be, and I agree with that, it shouldn’t be that “All right, I need a referral, who I can connect with so they can refer me.” It sounds a bit fishy and it’s going to come off ingenuine and might work out 1 or 2% of the time. But really I like your advice on build those mentor, mentee relationships. They’re going to serve you much better in the long run. You never know if they’ll turn into a referral they might, they might not. But as we discuss it, the start, right. It’s not about the goal, it’s about the journey. And along that journey you will learn tons and potentially, your end goal might change along the way as well. You might end up wanting to be in a different company or something else. So yeah, mentor, mentee relationship is really powerful.
Kirill Eremenko: We’ve talking about, we won’t go deep into this now because I’ve talked about this a couple of times in the podcast already and there’s a couple of other things I really want to interrogate you a bit more about. One of those things for instance, is your website. So guys, this is really cool. I highly recommend for people listening if you’re looking for a job in Data Science or if you are going in for interviews already, checkout Andrei’s website, which is simply andreilyskov.com. We’ll link to it in the show notes as well. And I love what you did here. It’s very simple, very straight to the point, talks about you. Has your photo, has your resume. So by the way, guys, if you want a really cool resume, I like it how it’s one page resume so you can get Andrei’s resume as a template, you can copy that.
Kirill Eremenko: And then you have your projects, you have some links to a code, dashboards and things like that. Really cool. In your description of the interview cycle that would really help you out with that whole HR screening. When they’re like, who is Andrei Lyskov and we type in your name in Google and this thing pops up, number one. And now they all of a sudden they know that you’re not just some person randomly applying. You’re pretty serious about what you do. You have some projects, you have a portfolio up there. Tell us about that. Like first of all, how did you come up with this idea to put everything up? And what triggered that for you? And also how has it helped you in your job application process?
Andrei Lyskov: Yeah, absolutely. The personal websites are very important for the exact reason you said where usually before people interview you, they’re going to Google you. And you want the first page results to be stuff that is about you and not random articles or random things that are not related to you. And obviously people have common names, it’s kind of hard for you to get top results in Google. Luckily mine is somewhat unique. And so usually if you type my name you will see like my site or my LinkedIn or some online presence I have. And so the personal site of mine actually has gone through a lot of different iterations. I remember the first one was like so complicated. There’s so many buttons and things and content. I’ve really embraced the minimalist approach where I think about how can I just get the bare content that I need to get across who I am and what I’ve done.
Andrei Lyskov: And so right now when you land on my site, if features my writing, because that’s what I’m focusing on and I want people to read and learn more about. Or as in previous iterations it might have like a biography of me and what I’ve done and my accomplishments and all this stuff. And that was necessary when I was looking for a job because I wanted people like recruiters or interviewers when they Google me, they want to see this run down. And so that’s what you would find on my website. Now that I have a job, my personal site has shifted in its purpose. Now it’s more focused on promoting my content, my writing, and building a readership. And so that website is actually a template as well from… I forget what the website is, but if you Google website templates, like html, CSS websites, you’ll find plenty of sites.
Andrei Lyskov: And I personally hate web development. I actually started programming because of web development. And I stopped because I was like, is this what programming is? Just html, CSS and JavaScript? I didn’t go back to programming for like two years because I was so demotivated by the boring this front end development. I have no qualms about purchasing a template and then editing it to my satisfaction. And for those listeners that are not CSS wizards have no fear, there’s plenty of templates you can buy online. And I think the one I have on my site might have been like 5 or $10. Definitely affordable and it’s really helped I think in interviews.
Andrei Lyskov: I actually had my manager at my internship where I was a Data Science intern talk about how he shared my site with his co-workers and got their thoughts on me before they made a decision and shared my projects with them. And that really helped I think when they were deciding on whether they would bring me on. Personal website, great investment, takes a weekend to put together if that, maybe even like a day. I would definitely invest in it if you’re looking to increase your chances of getting hired.
Kirill Eremenko: For sure. And some of these templates you don’t even have to code is just drag and drop. Very simple stuff.
Andrei Lyskov: Yeah. I know there’s like Squarespace and stuff, they have drag and drop templates. This one is actually totally like html CSS code and you just go into your editor and you edit the html and it updates. Still a little bit more intensive than say like a Squarespace or Wix or any number of ones. So yeah, if you don’t even want to touch any code, you can just do drag and drop and that is also perfectly fine.
Kirill Eremenko: Yeah. Even if you have to code a bit, you know why not, for a practice. Put that on your resume that you can code websites a bit as well.
Andrei Lyskov: Exactly.
Kirill Eremenko: Okay, cool. Cool, cool. And speaking of… I think that’s a good rundown of Data Science like interview process. And first of all thank you very much for sharing your journey. Is there anything you would add? Any I don’t know gems of insights that we haven’t touched upon that might help somebody out who’s about to go into an interviewers applying for jobs right now?
Andrei Lyskov: Yeah. The first thing to do is to define the role that you’re looking to get and break out the skills and attributes that consist of that role, particularly with Data Science because it’s such a broad field. As I had mentioned before, it’s really tough to prepare for a Data Science interview. And so Airbnb came out with a nice, heuristic or way to partition their data scientists, which was algorithms, analytics and inference. So with algorithms, these are data scientists that are typically, deeper into building models. They’re reading research papers, they’re implementing the latest and the greatest. Then you have analytics, which is more of the working with business stakeholders, SQL, some minor stats potentially. And then you have an inference which is much heavier on the stats.
Andrei Lyskov: And this is typically where folks with social science backgrounds will find themselves. So someone taking econ background might do very well in the inference Data Science category. If you’re just starting out looking for jobs, you think Data Science is cool and you want to learn more about it, I would really suggest you identify one of these areas and go deep because, I’ve certainly fallen in this trap when I was starting out. I signed up for all these different MOOCs, massively online open classrooms. I bought all these books and I was just drowning in content. And it wasn’t until I kind of broke out, what is my value proposition and how do I deepen it by preparing for that specific parts that I was able to land jobs and not feel like I’m drowning in content.
Kirill Eremenko: Great. Thank you. That’s wonderful advice. So indeed, Data Science is very broad. There’s lots of roles, from natural language processing to modelling to deep learning, reinforcement learning. There’s lots of different industries you could get into. Plenty of things. Define where you want to go first. And this actually brings us quite nicely to the next question I wanted to ask you. How did you teach yourself Data Science? Because you mentioned at some point that you we’re asking around what Data Science is with you or your friends and colleagues and how to get into this space. Why did you want to get into it and how did you teach yourself?
Andrei Lyskov: Yeah, I’ll answer the why first obviously, and then how I approach learning. At the time, as I was mentioning, I was a business analyst at Soylent and I felt like I wanted to go and develop deep expertise. I think at the time I had to read like deep work by Cal Newport and I was like, oh man in my role, I’m not really getting as much of that deep work. And so when I went back to school, I was studying computer science now instead of Bachelor of commerce that I was in originally. And as a result, I had this background of business and now I’m doing this computer science degree and I knew I didn’t want to completely throw away all the business background that I’ve built up. I didn’t want to be like a backend engineer, front end engineer or anything like that.
Andrei Lyskov: But there was this field of Data Science, which at the time… and I think now still is a nice intersection of a variety of disciplines, one of which is business. So being able to add business value, being able to communicate, work with business stakeholders. I felt like I had that skill developed whereas the technical was much less developed. And so once I identified that weakness I [inaudible 00:50:24] obviously was to simplifying that situation and deepening that expertise. And so in my case I taught myself through this deep immersion learning technique, which basically consists of trying to hit as many sources of knowledge as possible so that I’m surrounded by Data Science and data and programming.
Andrei Lyskov: I first and foremost was obviously taking computer science classes. So anything that had data in it, I would take it, statistics, databases, all that I would want to take that, take those classes. A second thing I did was projects. This is probably the best way I found to teach myself Data Science was building various projects. In my case I built a dashboard that visualized data for my life and allowed me to understand what is going on and allowed me to work with flask, built APIs, built a database, taught me a lot of different skills. Kaggle competitions of course they’re also a great way to learn through project based learning.
Andrei Lyskov: The second thing was basically online classrooms that were more focused on practice rather than theory. Dataquest for example, is an online web teaches you Data Science. So step by step, it’s an interactive development environment, very similar to Codeacademy where you have these different challenges and you go about completing them and working your way towards being more competent. Books. There’s a whole list of books I read. I actually have a list that you’ll be able to add in the show notes of books I read everything from a biography of Alan Turing to Weapons of Math Destruction, books on statistics, textbooks. It’s kind of whole gamut of books, podcasts, such as this, Data Skeptic, Learning Machine.
Andrei Lyskov: There’s so many resources. Communities. There was some slack communities that I was a part of where I could ask questions. And the last one was movies and TV shows. Shows like or movies, like Minority Report, or TV shows like Black Mirror, movies like Ex Machina. These were all ways that I got deep into the whole data science/machine learning space. And as a result, I was able to accelerate my learning because I’d be learning about specific context, specific topic in one domain, say a book. And then I’d also see it in a YouTube video. And this really helped, accelerate and reinforce my learning. And funny enough, I wrote about this, how I taught myself data science on Quora. I’ll share that link with you guys and you can dive deeper into that.
Kirill Eremenko: Wow. Fantastic. What a holistic approach to learning Data Science, I’ve never heard that one of about movies and TV shows. That’s very cool. What was your main takeaway from Black Mirror in terms of Data Science?
Andrei Lyskov: Oh yeah. So as you know the Black Mirror is the Dystopian things that can happen in the future if we abuse certain technologies, the shows or the episodes really we’re more about getting me to think about some of the implications of these algorithms that are built. Probably the most impactful one was where everybody was rating everyone in real-time…
Kirill Eremenko: That’s one of the first ones, right?
Andrei Lyskov: Yeah, exactly. That one was like, yeah, I can see it as happening and in fact it is happening in places like China.
Kirill Eremenko: Yeah, it’s already close. Yeah. Oh Man. I think I liked, they had a white Christmas. That was a really small blip for me. They’re all deep and dark, but that was cool. Or the one with the augmented reality horror stories that was also pretty insightful. They have a new show now on Netflix called “Love, death and robots”. Have you seen that one?
Andrei Lyskov: I heard about it. My friend was telling me but I haven’t had a chance to check it out. What is it about?
Andrei Lyskov: It’s like a similar?
Kirill Eremenko: Yeah, similar, but it’s shorter. So Black Mirror can go… seriously, those episodes can be like one and a half hours long [inaudible 00:55:02]. But Love, death and robots are shorter between 8 to maybe up to 20 minutes max, maybe like 12, 15 on average. And they’re all animated. They’re all like different types of animation. They’re all not related at all. So one might be a cartoon animation, one might be 3D animations. And yeah, similar concepts. They kind of pack that same idea as in Black Mirror, but into a shorter, more nicely presented, very quick rundown. And you’re like, “Oh, cool. I like that idea.” That one I don’t. If you don’t like it, you know you won’t spend 8 to 12 minutes on it, but they’re pretty cool some of them are also very deep. Nice show, also on Netflix.
Andrei Lyskov: Yeah, I’ll definitely check that out. I think part of watching this media is just like getting yourself psyched on Data Science and machine learning and being like, this is the stuff that can happen if over time people apply themselves in this direction. This is a vision that can be realized, some being dystopian or some being utopian. So you want to just see the breadth of ideas and for me a lot of these movies and shows just got me excited about participating in this field.
Kirill Eremenko: Yeah. And it’s interesting you mentioned that because when you watch them, for our listeners, try to… for any show you watch that has a technology, try to think of how is data working in the background? How is data working in that self-navigating spaceship or in that weapon that they’re using or in that space, facial recognition software, what role does data play? I think that can be a really cool way of watching these and getting more inspiration for real life as you say hopefully not dystopian but more utopian ideas that we can implement.
Andrei Lyskov: Yeah, absolutely.
Kirill Eremenko: Awesome. Well Andrei, thanks so much for coming on the show. This brings us to the end of the podcast being a great pleasure to chat and dive into your story. I’m sure it’s going to help a lot of people. Before I let you go, can you give us a quick, overview of locations online where people can follow you or get in touch for some more advice or just see where your career takes you from here?
Andrei Lyskov: Yeah, for sure. If you’re interested in reading some of my stuff online, you can follow me on Medium. Just my last name and first name. So Lyskov Andrei. And then I also write in Quora, I’m on LinkedIn all over the inter webs. So if you search for my name, you should have some stuff pop up, search, Andrei, whatever social media you’re looking at and something should come up.
Kirill Eremenko: Perfect. One final question for you today. We’ll of course link to all those in the show notes. And the final question is, do you have a book that you can recommend to our listeners that had the most impact on your career?
Andrei Lyskov: Yeah. I think with Data Science, again, because it’s such a broad field, there’s a lot of books and domains that one can read and recommend. In my case, the domain of Meta learning. Learning how to learn is perhaps the most important, not just for me, but I think for all data scientists because the field is growing so quickly and there’s so many new developments, you really do need to develop a love for learning and also implement strategies that are good for retaining information. So with that in mind, there’s this book I read a while back called Make it Stick, The Science of Successful Learning. And so this book basically outlines latest in research on effective learning and packs it in a really nice story telling methods that I really enjoyed.
Andrei Lyskov: And again, another bit of a plug I kind of wrote a post on medium talking about this learning how to learn and it’s basically a cliff notes of this book among other books I read. Make it Stick. That is the book I will stick with.
Kirill Eremenko: Very cool. Speaking of plugs, we didn’t have time to talk about your Data Minds series, but for our listeners out there, so Andrei is doing interviews with data scientists, which you might find interesting and they’re in written form as opposed to what you’re hearing now, a podcast. So you might want to check that out. We’ll also link to that in the show notes and sounds like you’ve got some very interesting guests coming onto Data Minds as well.
Andrei Lyskov: You can read more about it on medium.
Kirill Eremenko: Awesome. Okay, well on that note, Andrei, once again, thanks so much for coming on the show. Great chat and good luck with your new job at Apple. I’m sure you’re going to smash it and probably you’ll learn heaps while you’re there as well.
Andrei Lyskov: Yeah. Thank you very much. I found this to be great, really had fun time. Hopefully, I’ll have a chance to come back again maybe a year from now when I come through the reps of this new job.
Kirill Eremenko: Sounds like a plan. All right. Take care man.
Andrei Lyskov: Yeah, you too. Bye.
Kirill Eremenko: There you have it ladies and gentlemen, hope you enjoyed this episode of the SuperDataScience show and our conversation with Andrei and you got some valuable takeaways from here. For me, probably the most impactful was the trichotomy of control where things break into three categories where things that you can control, things at you somewhat can control and things that you can’t control at all and how to go about setting your personal and professional goals with that philosophy in mind. As always, you can get the show notes at www.www.superdatascience.com/261 where you’ll also get the URL to Andrei’s LinkedIn and other areas online where you can find him, other social media and websites where you can get in touch, connect and follow his career. There you have it. Best of luck with your interviews if you are going into them or with your job hunting, job searches.
Kirill Eremenko: Remember all the things that we talked about here and create a portfolio. Make sure that you are building your network, building connections and actually strengthening your online presence to make some of those steps that Andrei listed in his cycle of interviewing for Data Science and the whole Data Science recruiting process. You make sure that those steps are simplified for you and that your interviews are brief. On that note, thank you so much for being here. Make sure to share this podcast with anybody else who might be interviewing for a Data Science role or looking for a Data Science jobs. This might help them out and I’ll see you back here next time. Until then, happy analyzing.