Jeff Li

Jeff: [00:00:01] No, I'm doing this off the top of the dome, you see me, I'm definitely our pretz wearing that black sweater. You know what? There's coffee. You see me right here hitting on that line. Look at me. You know, in the there you see I see that birds in the sky outside of machine learning models, Data science. Oh, you see, I hit another rap session. Oh I've got a good ah. Seeing linear regression. No I'm doing this off the top of the dome. You see the name of Jeff Flake. Jeff Lee.

Harpreet: [00:00:51] What's up, everybody? Welcome to the artists of Data Science podcast, the only self development podcast for Data scientists. You're going to learn from and be inspired by the people, ideas and conversations that'll encourage creativity and innovation in yourself so that you can do the same for others. I also hosted Open Office Hours. You can register to attend by going to Bitly dot com forward. Slash a d. S o h. I look forward to seeing you all there. Let's ride this beat out into another awesome episode and don't forget to subscribe to the show and leave a five star review. Imagine yourself at a poker table, there's two thousand dollars in the pot. Our guest decides a Mullin victory is inevitable. He's got pocket aces, the best possible hand he could have at a poker table around the table. Before people start pushing all their chips in the middle, they [00:02:00] flip their cards and the dealer puts for community cards face up on the board. Mathematically, he knows he's got a 95 percent probability of winning the deal of drugs, the next card, and he loses like most of us and Data science. Our guest loves math but doesn't like losing money, looking for ways to do math and actually make money. He started exploring this data science thing. He's currently a data scientist at Spotify, and before that he spent two years as a data scientist at Door Dasch working on problems such as machine learning based variance reduction, demand forecasting and driver pay. So please help me in welcoming our guest today, an adventure writer and amateur freestyle rapper, Jeffrey Lee. Jeffrey, thank you so much for taking time out of your schedule to come on to the show today. Man, I really, really appreciate having you here.

Jeff: [00:02:50] Thanks so much. I appreciate it. Great introduction and excited to be here right on.

Harpreet: [00:02:54] And yeah, I'm definitely super, super pumped to chat with you. You're back. One of first things like launching the podcast, you're one of the first people I reached out to and didn't work out then. I'm glad it worked out now, man. So happy to have you here. Before we get into some of the awesome stuff that you write about in your blogs and some of the awesome work you do once you tell us a little bit about where you grew up and what was it like there?

Jeff: [00:03:17] Yeah, so I grew up in the Bay Area. I said pretty, pretty good childhood. Yeah. Kind of spent, say, grade school all the way to high school in a small town called Alameda, which is in the East Bay, very close to Oakland. And then I went to college in Los Angeles at USC and then kind of after USC broke into consulting, then I broke into Data.

Harpreet: [00:03:42] Science is my idea. I'm actually I'm born and raised in Sacramento, so I used to cut through Alameda right before getting on to the Bay Bridge. So, yeah, I know exactly where that's happened. So when when you're in high school, what did you think your future would look like?

Jeff: [00:03:59] So in high [00:04:00] school, I was obsessed with sports, so I wanted to become a professional athlete and I specifically I was really into American football, so I wanted to become a football player. But I knew as a five eight, one hundred and fifty pound Asian dude, skinny Asian dude, that career was probably not going to work out. But yeah, I'd say when I was a kid, professional athlete, I love sports. And then I soon realized that that wasn't feasible. So I had to kind of figure out something else to do.

Harpreet: [00:04:30] I got to know, man Raiders or Niners.

Jeff: [00:04:34] Definitely the Raiders. Yeah, I don't like the Niners.

Harpreet: [00:04:37] We might have to done this right now.

Jeff: [00:04:39] And and still, I don't understand.

Harpreet: [00:04:43] So talk to me about the journey that then led you from from high school, from a aspiring professional football player. That was your hopes and dreams to now becoming a data scientist. What was that journey like from there to now?

Jeff: [00:04:58] Yeah, I mean, in high school, I was always I always did really well in math. So that was often the subject. I would get A's. And so, yeah, probably from a natural perspective, I was a little bit better at math and other things. I think going into college, most 18 year olds don't really know what they're looking to do. So they're kind of partying, kind of roaming around. And then I think for me, I was not super like I think most people wasn't super sure what they wanted to do through high school. So I took the default route with where a lot of my peers went, which was consulting and pacifically technology consulting. And at the time, I got really interested in figuring out how we can learn things better and more effectively. So I was reading a lot about learning sciences. I was learning a lot about, hey, how can I like absorb information and have it stick better? And then the overarching goal was to basically teach myself how to learn things so I can apply it to like a wide variety of skills. So I started [00:06:00] experimenting on myself by, like, teaching myself freestyle rapping poker. That's kind of where you got the concept from the intro from. And kind of in the back of my mind, I was also trying to experiment to see what skills kind of resonated with me a little bit more and what resonated a little bit less. So after going trying poker, I didn't see poker as a good long term fit for me, but I, I really like the analytical side of poker. So that's actually when I started kind of initially self studying more of the programing and the Data science and that kind of fast forward now. It's almost four years. Yeah. Now now currently a senior data scientist at Spotify

Harpreet: [00:06:39] In this, by the way, that's like my dream job working on Spotify. Like they've got some amazing stuff that I won't touch on that later. But while we're discussing the topic at hand, you talk about learning how to learn and learning how to learn better. What were some of the resources you used to help you figure out how to become a better learner?

Jeff: [00:06:58] Yes, I think like most people who get into this stuff, they always start with like learning and memory techniques. So if you've ever heard of things like memory palaces, there's a really good book called Moonwalking with Einstein by Josh Foer, who is a journalist. And he what he did was he trained for about a year to become like a memory champion. And what they're doing is just like memorizing like long like hundreds of digits of pi, memorizing decks of cards within like a few minutes. I thought it was super cool. So that's how I initially got into learning sciences, realized it wasn't that applicable to my everyday life, because I then I soon realized that kind of developed skill development and not just memorizing information, but it kind of is more from deliberate practice. And that's why I started diving into a lot of Ericsson's work, understanding, hey, how do I structure practice? How do I set specific goals, set up a system to get feedback? And I think the [00:08:00] principles apply to any skill that you want to get better at. And then I got introduced to Scott Young's work, who is a blogger, and he did challenges like the MIT challenge, what you try to learn the entire computer science curriculum in a year. And he also did like the year without English, or he would learn four languages in a year without speaking English. So he has like he has a lot of great content on just how to structure your practice, how do you structure your learning? And then that's kind of how I got into it. It's a little bit of a zigzag. But yeah, I think that that is that's kind of those were the ideas that I toyed with at the time, and I still do use them today.

Harpreet: [00:08:37] In a recent piece, hundreds of groundbreaking research did actually. I interviewed Scott Young for the show a while back. I had an episode of that with him. And then there's somebody else that is also in that space. I'm not sure if you heard of Dr. Barbara Oakley. She's had that course learning how to learn on Coursera. She also wrote the Book of Mind for numbers. So we would probably dig her stuff you have interviewed. As well, the episodes going be releasing at some point, probably before this one, but I'll definitely link you to that stuff, man. So I love your writing, man. I got a lot of awesome blog posts and a lot of awesome content. And speaking about deliberate practice, I love this remix you did with the ideas of deliberate practice and Data science career advancement. So how can we use deliberate practice to improve our skill in Data science? I guess just for whoever's listening? Can you kind of define what deliberate practice means? And then how do we apply that to improve our skill in Data science?

Jeff: [00:09:33] Yeah, so at a high level, deliberate practice is the process of. Practicing where you set specific goals and targets for each one of your practice sessions and ideally you have a coach there to give you immediate feedback that way, the kind of feedback loops kind of will increase the speed in which [00:10:00] you will improve at that specific skill. And then specifically, when you are practicing, you are only focused not only, but for the most part, focusing on the areas where you're the weakest that. So, yeah, that's probably the high level definition I can come up with.

Harpreet: [00:10:18] Yeah, that's definitely very when I first heard that that concept, that idea was pretty groundbreaking to me, that there's actually a specific way that you should be practicing the things that you do. One thing that I've been doing is trying to up my typing speed. And I found that there's some software out there. I forget the name of the website. It's been a while since I worked on it. But that's you know, you need measurable outcomes in order to improve with deliberate practice words per minute. For example, this is one such metric. But how do we do that for career advancement to improve our skills in data science? How do we apply that concept?

Jeff: [00:10:54] Yeah, so it starts to get a little bit trickier because a lot of the studies that Ericsson has done is with chess players and chess players. Like when you're playing chess, it's a predefined set of rules. And if you read the book by about generalists, they I forgot the name Brain. Yeah. Range. He talks about David Epstein. That's right. David AIs talks about this idea of wicked games where you're playing games where the rules aren't are kind of dynamic and they're not clearly defined. And I actually think when you're learning Data science, it's a little bit more of a wicked game than something more like chess or golf. So I do think it's a little bit more difficult to apply deliberate practice to it. But I generally like your original question was how do you apply deliberate practices Data science. Right?

Harpreet: [00:11:46] Yeah. To advance in your career or improve your skill rather. Yeah.

Jeff: [00:11:51] Yes. I mean it's like really tricky because Data science is such a big field and there are so many different topics that you can learn about at [00:12:00] a certain point. It's really like impossible that you have deep learning networks, causal inference like reinforcement learning, and people are getting like these in like a nascent topic within, say, like NLP or deep learning. So there's definitely always more to learn and like more to get better at. So I think that, like with Data science, the key is actually figuring out like the end goal first. Where do you want to end up within Data science? Do you want to become a superstar researcher? Do you want to become like a machine learning engineer? Do you want to start your own company? And then once you kind of clearly defined exactly the outcome that you're aiming for, it makes it a little bit easier to start breaking it down into chunks. Of course, that getting with getting that outcome is really, really hard. I actually don't really know my personal outcome, like what my kind of long term goal is with data science yet. But I do feel that once that is very clear, like, let's say you wanted to start your own, like, machine learning based company, then it starts to be clear on exactly the roadmap you have to do to get there.

Jeff: [00:13:06] So, for example, if you want to start your own small based company, it probably makes more sense to join a startup like a machine learning based startup than to go work for like a giant company like Google. I think both would be great experiences, but working at a startup is probably a little bit more in line with your long term future. And then you just continuously work backwards and continue to break it down, say, OK, what skills does a startup value? Well, they probably are not as developed on the engineering side, specifically for like deploying models, whereas, say, like a bigger company is probably much has much more developed infrastructure. So you don't have to be thinking about, say, tracking events or making sure the right infrastructure is in place to do analysis. Therefore, you probably need to know some more engineering skills and that can kind of lead you to the resources that you want to really hone in on

Harpreet: [00:13:58] All of them. And thank you very much for for diving [00:14:00] deep on that. And within this realm of deliberate practice, there is a undertaking that you had taken. That was the month to master project talk to us about this.

Jeff: [00:14:12] Yeah. So I was honestly very inspired by a lot of the work Scott Young has done. So I wanted to come up with my own and say, like ambitious, really difficult learning project on my own. I saw another guy named Max Deutche. He kind of did something similar. And this is an idea I was always toying with wanting to do, but I just never did. It didn't think it was actually possible. So once I saw somebody else do it, it like opened up in my mind. I basically was like, oh yeah, this is actually very, very possible. I could I could actually do this. So, yeah, the overall goal was really to just kind of really understand the process of developing skills and just to see how well I can how good I can get at a specific skill within a month. And I think at a larger level, having such a large goal does bring a little bit more purpose to my daily life because I have such a large goal. I know what I have to do every single day to get there. So I would say that, yeah, that's kind of the high level kind of thinking behind the project. And then while going through it, yeah, just I think my other intention was, hey, if I can, like, inspire somebody else to. Pick up a new skill that they're interested in or they might find useful that that would say that would be a win.

Harpreet: [00:15:23] I like the man like to know the story about Roger Bannister. He's like the first guy that was minimal and then show the world as possible. So you kind of had one of those moments where you saw somebody achieve it, like, oh, shit, I could probably theCIA myself and then just went for it. It's awesome, man. Definitely. Guys, I'll link to your blog because you've got a lot of wonderful articles on there. Really, really enjoy your writing style than most got the article or page or whatever on on your website. That's just entirely about mental models. And I love how you say that by installing different models in your mind, you end [00:16:00] up being better able to understand the world and to kind of see it through a different lens. So first of all, I guess, how would you rather define a mental model? And can you describe maybe two mental models that you find yourself having to deploy most often in your work as a data scientist?

Jeff: [00:16:18] Yeah, it's a great question. So I would say for mental models as a whole, yeah. The basic idea is like learning ideas that kind of, you know, either like what? Stand with stand the test of time and learning ideas that you can actually, like, pull from a very wide variety of fields and apply it to your own life. So I would say that in general, the mental models that I use the most as a data scientist, I'm actually using a lot of the mental models that I'm going through right now. So I actually see a B testing as kind of a big mental model that I use all the time, I think. And it not just applies to just running experiments for your job, like testing new features, but you can actually apply it to everything. So one example is, let's say I wanted to write. I was thinking about writing a blog post. Well, I don't even have to write a blog post first. What I can do is say, hey, I have this idea for a post. I'm actually going to try to bring it up in conversations with friends to just see how they react to that type of idea. And in a way, I'm actually testing the receptivity of that idea with a friend to kind of validate the idea. And that's actually how I came to the conclusion. That's how I came to write that post on how I got seven offers.

Jeff: [00:17:35] And during the pandemic is because a lot of my friends, they were actually really interested. And like when I started talking about the statistics of my job search, they always keep asking me, oh, like I said, you know how many applications that you have to send out. And like like, OK, maybe there's an appetite for a more analytical description to job hunting. So that's one that I am always using, like I'm using constantly. One outside of Data [00:18:00] science. I definitely apply to my life is this idea of activation energy and it comes from chemistry where you need to like hit like some minimum threshold for like the chemical, I don't know, chemistry that well. But you need to hit some minimum threshold in order for a chemical reaction to occur. So I find that very applicable to like, say, when I am feeling lazy that day and I need to go exercise. I know that I don't push myself to do like a one hour workout, but I just tell myself, OK, I'm just going to do like a small walk. And then usually once I convinced myself to do a small walk, I hit like a tipping point. And I'm actually willing to do it like a full workout. So I would say those are like two that I am kind of using in my day to day.

Harpreet: [00:18:44] I like that a lot. There's a couple of writers that I really admire, like James, Alicia and Screengrab. They do something similar to this that you talk about, talking to your friends before writing a blog post. They'll start with like a Facebook post or something like that that they'll see with the reception they get on. That is. And if the results indicate that they're getting some good reception on that, they'll go write the full article on that. And I like that idea of activation energy. I think there's maybe was an atomic habits or tiny habits where he's talking about just told somebody just to just do this, just drive to the gym everyday. You don't even have to go inside, just drive to the gym everyday, make that your thing, and then eventually you'll get to a point where you're like, oh, shit, I'm already here. I might as well as well workout or whatever. Yeah. So which mental model would you say has had the biggest positive impact on the way that you see the world.

Jeff: [00:19:43] Yeah, I would say right now the one that I've, I felt that was the most impactful is this idea of probabilistic decision making. So like I kind of this is kind of a newer mental model for me. So just [00:20:00] to give you some context, there is a mathematician named Enrico Fermi, and he invented this thing called Fermi Estimation, where the basic idea is he tries to estimate it's like the standard Google, like how many golf balls fit in a school bus. It's like basically a method to try to answer those type. Have questions so you can actually make very informed guesstimates about what's going on in the world. So I found that to be this kind of model, I would also like this is kind of like probabilistic thinking. I found that this model's really useful for kind of understanding the different decisions or the kind of analyzing the different events that might or might not happen in my life. So I think one example is like if I'm trying to make decide whether I want to want to start a business or not. Well, I think applying this type of thinking to it, well, you can basically try to you just basically want to size out what the total addressable opportunity is, whether people are willing to pay for whatever you're offering and actually computing guesses you're in a way like a predictive model trying to predict how much revenue you might make. So I found that that model to be very useful in just helping me make better decisions and kind of really understand, OK, this is kind of the magnitude of impact for this specific decision. Yeah, a little bit of a scattered answer. I never really articulated that. Yeah, definitely. I didn't think about that. But yeah, that's what comes to mind,

Harpreet: [00:21:27] I tell you. Makes sense to me, man, actually that thinking about the future as a probability distribution kind of along the same lines which you're talk about, which is viewing the future as just a probability distribution, has completely changed my life. I first read about this idea, this concept a few years ago in thinking and bets by any do and that just like rattled me to my core, just that way of thinking it was insane because the success equation by Michael Globsyn, and it's along the same lines about what we're talking about here. So I think you might definitely enjoy that to [00:22:00] talk to us about your problem centric approach towards doing Data science work. First of all, what does this mean, problem centric approach? And then walk us through this four step formula that you have to solve like any problem.

Jeff: [00:22:13] Yeah. So problems centric approach to Data science is where you focus on the problem that you're aiming to solve and you're agnostic of the tool that is supposed to solve that specific problem. So I think in Data science, a lot of times everybody falls in love with deep learning or machine learning, when in reality sometimes a simple heuristic or just a quick exploratory analysis can actually solve that problem. I see this most often with people who are a little bit more new and junior to the field, and it makes sense. I was like this to where we learn about all these machine learning models and we just want to apply it to everything so we get a take home assignment and we just start fitting models on top of it. But then when I get questions or if I ask questions saying, OK, OK, we have this model, what can we do with it? Like, how can we make this useful for the business? Usually candidates have trouble answering this type of question. So it's a good example of focusing too much on the tool and not enough on the problem. If we like if they just focused on the problem, they could have like fitting models usually takes it can take a lot longer and you might have to wait a lot longer. So sometimes you want to find the trade off between speed and just getting that problem solved. The answer your second question, kind of the four step approach to solving any problem that comes from an article I wrote a while back. I would say that I think my approach is kind of a little bit more involved than the approach that I have mapped out on that blog post.

Jeff: [00:23:50] But I generally feel like the first step of any problem is to just understand the problem. In the article I talk about, hey, we want to make sure we actually understand [00:24:00] one, what exactly we're trying to solve or to what the constraints are. And when I'll add on top of it now is we want to be we want to break it down where each individual component is independent of each other. So this concept is actually very popular in management consulting, where you want to be mutually exclusive, collectively exhaustive. And that way you can isolate each component of the problem and solve each one of those. Second step is usually to devise a solution, and the way I usually approach this now is I will actually devise multiple solutions and I'll actually rank them and prioritize them based off of my guesstimate, applying kind of probabilistic thinking on what which solution I think is going to going to work the best also against how difficult is this and how complex or simple it is. So we also want to apply Occam's razor, which is, you know, go with the simplest solution if both are equal. And then the third and the last step is to just like validate evaluate the solution or like execute the solution. Then you want to evaluate the solution. So those first two steps, I would say, are more important. It's kind of like sharpening the knife. And then once you have that knife sharpened, you can actually just slice through that tree. So, yes, that's kind of like how I would describe that four step approach to problem solving.

Harpreet: [00:25:19] Thank you so much for sharing that. I know a lot of people are going to really benefit from that and probably rewind and listen to that a few times. And that's very, very valuable. It's no wonder that you ended up with seven job offers during the worst job market in history. Man, with with a mind like that. It is. It is. No wonder that happened. So dig into this man you're talking about. You got seven job offers in this climate. What is the secret? My man, how did you do this?

Jeff: [00:25:47] Yeah, I don't know if I would say. I'd like to troll people, so I usually just tell them it's magic, but that wouldn't really make a good fit for a podcast, but I would say. What is the secret to getting seven job [00:26:00] offers? So I think that there's a big question. I would say there's a couple of different things that come to mind. I would say the first thing is there's no secret, but I would say there are better strategies that you can use to increase your chances of getting a job offer. I think that the first thing is I am I was in a lucky position where I did have a few years of experience. So it did help that I did have some experience. If you read the article based off my analysis, I think that I had to I averaged about one offer every 10 applications. But when I first started, it was about one offer every 60 applications. So it does, I think, for people who are, you know, have a little bit more years of experience, just know that it does get easier as you get more work experience. I think, too. I have a little bit more of an unorthodox way to getting interviews where my thinking is I usually try not try to think about what everybody else is doing. And then I try to think of, OK, if everybody else is doing this, like doing X, how what can I do to stand out from the crowd? Because I don't have a traditional background. I didn't get a math PhD I don't even have a math degree for undergrad. So I would say that, OK, well, I can't compete just like through that front door.

Jeff: [00:27:24] I need to figure out like another door to go through. So usually my approach is to actually I mean, referrals are clearly the best, the best avenue for getting interviews. But if you don't if that's not available to actually find hiring managers and recruiters and write them cold emails and these emails actually have to be good, I think that good meaning personalized, you kind of showcase your skill set very well and you have links to actual evidence that you've made impact and have the skills to perform the job effectively. So I would say that that's one strategy I employed quite a bit. And [00:28:00] then the second I would say the third strategy I tell people is that for jobs that you really want, make sure you ask good questions throughout the interviews. Once you get the interview specifically, you want to ask questions like, hey, what are the biggest problems or challenges that you face from your perspective? And it might just sound like a random question, but you actually want to take notes on their answers here, because once you understand their problems, then that gives you kind of a formula to come up with ideas or kind of maybe build a project or add value to that company in a way that no other candidate is actually doing. And if you can actually prove that you can add value without any context into the company, you're definitely like a clear hire from the hiring managers perspective. So I would say, like, those are the three things that I would say, like kind of a little bit more unorthodox that I did that most people won't do.

Harpreet: [00:28:51] And I absolutely love that doing things that other people aren't doing. Right. So you can control every aspect of the job search process that is under your control. You can't control whether or not somebody actually gives you a job offer or whether or not you know it. There's so many things that are outside of control. Just don't even worry about those with focus on the things that you can't control. One of the things that you can control is this completely falling in love with the company, going on their blog, learning about their culture, learning about the work that they do, understanding their products, even buying the product and playing around with it just to understand it, if possible. Because if there's like a hundred people trying to apply to that exact same job and they all look just like you on paper, then your technical skills aren't what separates you. Right? It's the things that other people are doing, like everything you were. So I think that's a great way to approach it. Thank you so much for sharing that. So I know that me personally, when I'm out there applying for jobs, I can get sometimes so emotionally invested in just any one given prospect. Did you ever feel that at all? And if so, how did you deal with that?

Jeff: [00:29:59] Yeah, so [00:30:00] I mean, I think for a company I really wanted to work for, I definitely it's easy to get invested in one prospect. It's actually very similar to dating where you're really invested in one person you're looking to date. So it's very, very there's a lot of parallels there. I think that one thing that I found to help alleviate that is to actually just keep multiple balls in the air. So, like, I know that I'm really interested in this one company, but I need to still if I kind of diversify that kind of mental investment across multiple companies, that allows me to just not be as hung up. If I didn't get the offer or if I got rejected from that company, so ideally, if you can have like three, maybe at least maybe two companies you're talking to at the same time, then that way you're you're not you're like emotionally not just overly invested in the decision of one company, because it really is it does come down to a bit of luck, like you need to hope the competition is also worse than you. They're not better than you. You also want to make sure it's a good fit, unlike all across all five people that do interview you. So, yeah, I would say, like have multiple balls in the air that I found that to be a pretty good solution. And yeah.

Harpreet: [00:31:19] Hey, are you an aspiring Data scientist struggling to break into the field or then check out DTG dot com forward slash artists to reserve your spot for a free informational webinar on how you can break into the field? It's going to be filled with amazing tips that are specifically designed to help you land your first job. Check it out. DSD Jay dot com forward slash artist. So what advice would you say can you share with people who are just looking at these job descriptions and some of them look like they just want the abilities of like an entire team? Right. And so they might just feel scared [00:32:00] or dejected and not applying for that or get discouraged from applying. What do you got to say to somebody, that feeling that.

Jeff: [00:32:09] Yeah, so I never met all the the job requirements when I did apply. I think that the my general rule of thumb is if I matched 50 percent of those 50 to 70 percent, I would still apply. I would apply for it if at like if I really only matched like twenty five percent of it or less, it's a little bit of a stretch because I think usually, you know, they don't require you to have everything on their job description. But I think there are some core components of it. So I would also think about what looks like is essential in this job and what looks like is kind of a nice to have. Like, I think for a lot of machine learning engineering positions, it's like you have to know how to to write code. That's like if you don't know how to do that, you're you can't do the job. But if it's like some kind of nascent library, such as like, oh, we want to hire data scientists, but we would like them to know how to use like deep learning algorithms or whatever. I think deep learning is usually a nice to have unless it's like core to the product. So like if they say if they say something like that, you can probably apply for it because you can you can learn on the job. I've been close to getting offers from machine learning engineering positions specifically for deep learning and I don't really know deep learning that well, but they were confident that I could learn it on the job. So I think that you should also differentiate what is essential and what is more of a nice to have.

Harpreet: [00:33:36] So you went on so you got seven job offers. So if I buy from using your math right, that means you went on at least 70, 70 interviews this year, 70 applications.

Jeff: [00:33:47] Oh, yeah. Yeah, probably like 70. Yeah. Quite a bit of interviews. Yes. OK, interviews

Harpreet: [00:33:51] Seven. Oh yeah. That's right. Because you're saying ten applications equals to to one to an hour regardless. You went on a shit ton of fucking interviews [00:34:00] in twenty twenty man. So. What was common among all of these interviews, like what themes did you see kind of bubble up to the top in terms of, you know, I mean, I'm sure you got tossed a bunch of hacker ink or leaked code type of assignments. What are some key things from that experience that you think that people really need to go go focus on?

Jeff: [00:34:25] So the common themes that I've gotten from all my interviews. So I did kind of do an analysis looking at question categories, not as didn't get as many interesting insights, I think that the core takeaway is that with Data science, you're just going to get a wider range of questions than other fields so you can get asked SQL Python experimentation and modeling. I think the core themes that I've seen are that one school is pretty much like table stakes. You definitely need to know that if you don't know it, it's depending on the role. It's usually like it's automatic. No. Another one is like the rise of machine learning case studies. So I would say a lot of time, a lot of questions now are they give you a business problem and you try to use machine learning to solve it versus the first time around? When I went through the interview, I wasn't didn't really see many of those. So I think getting air tight on not just your machine learning knowledge, but understanding how to apply different concepts to different business problems is actually more and more important. So I would say those are the two things that stood out. And I think the third one is knowing how to talk about your projects really, really well. I think I've gotten to a point where it was automatic or I just didn't even need to think about it. I could just start talking about my projects. But [00:36:00] there's like a lot of interviews now where you talk about projects. It's not just like talking about your project, but they get into the technical details to see if you actually worked on it or not. So, yeah, I would say those are the three things that I glean from all the interviews. But I think my first point, there's a wide variety of questions. So it's a little harder to get kind of. Yeah, there's a lot of variance.

Harpreet: [00:36:21] Yeah, I definitely think the audience is going to appreciate that. Some good takeaways from that. Thank you for sharing that. When it comes to sequel, there's obviously there's the basic stuff that I think, you know, everybody should know the big six classes. What other nuanced things do you think a Data scientist should really, really understand from school?

Jeff: [00:36:41] So I would say you have a lot of people have asked me about how good you need to be at school. I would just say that if you can answer like medium, medium question, medium to hard questions on hacker rank, you should probably be in a good spot, I would say. I mean, I think a lot of it is just doing a lot of practice questions. I would say that one tip I would give is, is to actually write out your logic and comments first before you start writing the school code. I found that talking about your thought process to an interview is often really, really useful in that they can actually, if you write out the logic, they can easily follow your logic and if they can follow your logic, if you make a small mistake, then they can actually help you and correct you and push you onto the right track. However, if they cannot follow your logic, you basically have to get the right answer. Otherwise, it's like, hey, I don't really know what you're doing. I actually think this applies not just to school, but any coding problem where communicating it in a way where the interviewer can understand your logic that opens the door for them to actually help you out more and help you pass the interview.

Harpreet: [00:37:52] Right on, man. Thank you so much. I know the audience is really, really good. Appreciate your insights on that. So I guess we touched a little bit about [00:38:00] people who might come across these job descriptions that look crazy and get discouraged from applying. And sometimes somebody will see this job description that looks crazy will apply. We'll get through the entire interview process and get an offer and then end up having to work as a data scientist, which could sometimes lead to feelings of imposter syndrome. So what is your relationship with imposter syndrome like?

Jeff: [00:38:28] Yeah, I mean, I definitely still experience it from time to time. Yeah. When I was switching jobs, I got promoted. I was like, oh, like I was a little bit not surprised but like also a little bit more like intimidated. I'm like, OK, like I need to like kind of really step up my game. So a little bit. There is definitely a little bit of imposter syndrome creeping up. So there's a couple of things that come to mind. I think that one is that people who are experienced are Googling like really, really easy, easy shit, like I think the other day I forgot how to append a row to a panda's Data frame. And it's like, I don't remember how to do this, so I had to Google it. So people who are more experienced, who are Googling really, really easy things. That's like one that's like a good reminder that, hey, like, you know, it's nobody actually knows, like everything I would say. The second one is that basically like the field is so large that once you, like, gotten down two or three projects within a certain area, you probably know more about it than a majority of other data scientists. So if you work on causal causal inference for like about three projects, you probably know it better than most Data scientists. So even like more experience data scientist. So I would say that the key generally imposter syndrome is actually to use it to keep getting better. I think that not having like having a little bit of imposter syndrome is good [00:40:00] because it forces you to say, hey, I don't think I'm good enough, I need to get better. Having too much of it can be crippling, but having a little bit what can actually motivate you and push you to continue to grow your skill set, which ultimately is beneficial for you and beneficial for everybody around you.

Harpreet: [00:40:15] And I agree with you. I mean, the people get imposter syndrome because they are afraid of something. And, you know, fear is a powerful, powerful emotion. If you can wrestle it and have a push you from behind rather than block, you stand in front of you, then. Yeah. Have they push you from behind how you learn whatever it is you've got to learn and improve and grow. It could be a very powerful force in that respect. Thanks very much for sharing them. And any tips or words of encouragement for people, even if they're trying to come back through an imposter syndrome? Do you want to give a quick pick me up?

Jeff: [00:40:50] Yeah. I mean, I think kind of reiterate what I was saying earlier. The most experienced Data scientists are Googling the easiest things. So and I would say that technology is just changing. There's always kind of a new algorithm out every five to seven years. So at some point, people who are much more experienced are going to just be on the same playing field as you are. Granted that, yeah, they do have more experience so they know how to tackle problems like a little bit better. But just know that because the industry is changing so much, that levels the playing field quite a bit. And I would say the last thing is just to like, never stop learning, just continue to keep learning, staying curious, keep aiming to make yourself better, because if you just keep making to make yourself better every single day, I think once, like three hundred sixty five days passes, you'll see that you're way, way better than you were last year. And that's that's how I feel when I kind of look at my skill set a year ago and two years ago,

Harpreet: [00:41:49] One percent every day. That's like what? Thirty eight x improvement over the course of the year? Something like that. So I'm just a little a little bit steps at a time goes a [00:42:00] long way. So you've got some really, really cool, really creative Data science projects. I'm curious man. Talk to us about the importance of having a portfolio project. If you don't have any experience working as head of scientist, do you think that portfolio projects are enough to kind of bridge the gap?

Jeff: [00:42:19] Yeah, I mean, I think that so, yeah. Two thoughts there. So I would say that portfolio projects are good showing employers that, hey, I have the skills, I can write code, I can go. And and with the Data science process, however, I would add that just purely building your projects is not enough anymore because everybody is building portfolio projects similar to this theme. If you want to kind of you don't want to overly focus on competition, I think this is also like a business lesson. But you also want to kind of keep tabs on what's happening in the market. If everybody has a portfolio project, then your portfolio projects, by definition, will not stand out as much if you're doing the same projects as everybody else's. So you're doing like Titanic or like that. They're just Data set like everybody does that it's not going to stand out. I think what makes a portfolio project stand out more is, one, if you actually build something that solves an actual problem in the world, that and I actually can give you like a seed to start a business at some point, but holding a project that solves a specific problem that can maybe be used as a webapp or that can be used by somebody else. I think the third component is actually turning it into a blog post and that can actually showcase a lot of your communication skills. When I like ask a lot of other data scientists their biggest need, that kind of reason why people fail their interviews is. Because they don't have good communication skills, like a lot of people know the machine learning algorithms, but not being able to explain it [00:44:00] well is very, very important. So I actually would say turning it into a blog post is actually can help you stand out a little bit more because it forces you to showcase your writing ability as well. So to conclude, actually, portfolio projects are important, but like you need more than that as well to really stand out from everybody else.

Harpreet: [00:44:19] What are some of the biggest mistakes or biggest journalistic mistakes? What are some of the biggest mistakes you've seen? People breaking into the field make with their projects.

Jeff: [00:44:31] Biggest mistakes, I would say, I touched on this earlier, building a model when it's not really clear how the model solves a problem, that's the first one just fitting a model and everything. Second one is, yeah, doing projects that everybody else does, which is the minute it's called I to pronounce it the digits data set and Titanic. And then the third one is. Yeah, I would say those are the top two that come to mind immediately. And then I think yeah, the third one is probably just applying directly to job boards. That's like, you know, you might get some replies there if you're just breaking in. But, you know, you're competing against hundreds of thousands of other Masters students who just graduated as well. So so my general theme is just figure out ways to stand out from the crowd.

Harpreet: [00:45:21] So I've definitely got to get into a couple of projects that you've done because I found them to be so freaking awesome, so creative. And look, before you get that, just you know, I'm curious when it comes to standing out in the job search process, standing on the interview, I think one big way for you to do that is by having a well crafted story, well crafted response to the Tell Me About Yourself question. And it's surprising to me how many people don't know how to answer that question. So I'm wondering if we could do a little role playing here real quick. We can pretend that we're in an interview and and ask you, Jeffrey, tell me, how would you answer that question?

Jeff: [00:45:58] Sure. Yeah, it's all on the spot. I haven't done this [00:46:00] in a while, but. So tell me about myself. So so I'm Jeff Bennett, Data scientist for about three or four years, initially got into Data science transition from technology consulting, broke into Data science by actually trying to become a semiprofessional poker player, realized that wasn't a great fit. So that's kind of how I shifted to the Data science. First job as a data scientist was that Data quest where I was actually writing courses in Pythonic Machine Learning, where I built out, where I really kind of honed my communication and both written and verbal skills. I moved to Jordache, focused on more small problems, started off working on recommendations, then moved to projects such as demand for castings, multi touch attribution for marketing, and then spent over a year on the logistics side of the business answering questions on like how do we fairly pay delivery drivers? How do we make sure demand and supply is balanced? And how can we actually improve our experimentation capabilities? Move to Spotify. And now I'm in charge of leading to projects where the first project is building an end to and adds demand forecasting model. And the second project is doing fundamental research into figuring out how we can better monetize podcasts. And yeah, that's kind of my my background.

Harpreet: [00:47:20] I love that nobody has taken taking notes that it's an excellent way to answer that question. Speaking about research on the monetizing podcast, you share some of that with me. So I paid for this every day on my Spotify is awesome. Do you like it? That's like my dream company to work for. I think the Data you guys have like part of the the song metrics API and stuff like that. Super, super fascinating. I'm curious if somebody wanted to do a project using data from Spotify, what do you think would be a good project idea?

Jeff: [00:47:55] So I think a few ideas come to mind, actually. One [00:48:00] idea I was toying with I'll just give it away for free is basically to figure out how to how to match my playlists to my beats per minute, basically build a recommendation system based off of what my heart rate's like, basically. So then what it can allow me to do is connect the ideal song for which workout or like what level of intensity of a workout that I'm working on that I'm doing. So that's that's like one project that I was trying to do. I would say another one could be. I guess the first ones that come to mind are recommendation systems of trying to think of other ones that yeah, I'm also thinking about it from, like the advice I gave earlier about solving specific problems. So perhaps I could kind of turn the Spotify Data into some type of like marketing sizing engine to kind of figure out, hey, if I wanted to become an artist, what is the size of the fan base and what kind of targets? What I need to hit to eventually be able to go viral. I think I would also probably be interesting to understand reality like kind of relative to the rest of the world in the news, if I'm launching songs like How many songs do I need to launch until I like her hit a certain fan base? I think this applies to YouTube and podcasting as well. But yeah, I'd have to think about that a little bit more. Haven't played with the API too much, so I don't know exactly what data is available.

Harpreet: [00:49:24] But Gazmin, thank you so much for sharing that. So talk about how you use deep learning to help you find dates here.

Jeff: [00:49:33] So I think a few years ago I was wanted to learn more about deep learning and convolutional neural networks and I didn't really like using dating apps because I just felt like it was a waste of time swiping through mindlessly, swiping through things, through profiles. So I thought that what I could do is build a model that would automatically classify profiles for me. [00:50:00] So I did that. I scraped images off of both the Web and using Tinder. I used a convolutional neural network. I first used a three layer network to train the wave and I pulled the last two layers of that three layer network. Then I took like a pre built model like BGT 19 from I can Google and then I just replace those last two layers with the weights from the model that I trained. And then I hooked it up to the Tinder API so I can actually swipe through. I can set a set the number of swipes that I want to make per day. And then the script would automatically run. And I added some functionality where I can add like an automated message. So if I did match, it would just automatically send a message for me.

Harpreet: [00:50:51] That's a project. And so it's still out there looking for the one.

Jeff: [00:50:58] Yeah. Still right now. So I think that yeah, that was a few years ago. So I stopped actually stopped using using the algorithm because I started using other dating apps. I had to kind of figure out how to access their API. So I just haven't done the work yet.

Harpreet: [00:51:12] Ladies, if you're listening, he's single. I mean, it's mostly dudes, so.

Jeff: [00:51:18] Yeah, usually more data science. Yeah.

Harpreet: [00:51:21] So what are some non obvious skills that Data scientists are missing that you think that they should go pick up and maybe how can they cultivate those skills? Because I mean obviously creativity is one of them. You've definitely got that covered with the projects in writing and all that stuff. Yeah, well some don't obviously think that we need to be successful in this field.

Jeff: [00:51:41] Yeah. I mean, you actually kind of touch on that with what you just said. I would actually say writing is very, very important as a data scientist. I think that because a lot of times when you're building at Data science projects, you're not just writing code from the beginning. You're trying to scope out the problem and kind [00:52:00] of create a roadmap as to how you want to tackle the problem. And that actually comes from clear writing. If you're going to clearly write, then you can have a clear train of thought into how you're thinking about the problem. And other people can understand your train of thought on how you're thinking about the problem. So I would say I think communication I wouldn't I would say communication is very important, but I think everybody knows that it's important. But writing is one skill that is not really as glamorous within the Data science space. But I actually think it's very, very important, especially as you get more and more senior. You're not going to be doing as much coding, but you would be doing more like, say, planning, organizing, figuring out what is the right thing to be working on

Harpreet: [00:52:45] A hundred percent of with them and like pick up a quick business writing course on LinkedIn. There's plenty of them that are just like an hour or two hours long. We're. Check out some videos by Scott Adams, creator of Dilbert. He's got a couple of sessions on YouTube that are just all about how to become a better writer. Do you keep a journal or anything like that?

Jeff: [00:53:03] Yeah, I mean, not like a physical journal. I use RIM research, so I use room to write my thoughts down. And then sometimes when I do feel like I have something I need to articulate, such as like after getting a good amount of job offers, I felt like I had some something to say and some insights to share. So when I do feel like there's I have something to share, then then I'll actually try to make the ideas much more.

Harpreet: [00:53:26] Yeah. I write in for journals that I write in the morning and to the point about how writing helps you clarify your thoughts. Just clarify thinking like after writing my journals in the morning, I just feel like stillness in my mind. There's no chaos anymore. He's like, dump all this shit out, clear stuff out from your mind. It's just like a nice, tranquil stillness that lasts for a little bit then the days.

Jeff: [00:53:53] Yeah, yeah. I think it's like a lot of times when there's something on my mind or I can't really sleep at night or something kind of bothering [00:54:00] me when I, when I just write it out, it kind of separates me from that feeling or emotion and actually just feel a lot better. Yeah. I think writing can be used for like kind of expressing your emotions or feelings or you can if you need to think through something. I found that putting it on paper really, really helps as well.

Harpreet: [00:54:19] Speaking about communicating, what tips can you share with Data scientists for developing their leadership and influence skills?

Jeff: [00:54:28] Yes, I mean, I think I'm still trying to develop that myself as well. So I'm not sure if I'm super qualified on this, but I can give it a shot. I would say that in terms of, like leadership, the one you mentioned was communication or what was the second one you mentioned?

Harpreet: [00:54:43] Influence and leadership.

Jeff: [00:54:45] Yeah, leadership and influence. So one lesson I learned from being adored door was that as you provide more value and you're reliable and you're providing good work, your influence will gradually grow, mainly because if you're doing good work, if you're solving problems, if what you're doing is making an impact, people will see that and you will have you will start to develop expertize and knowledge and then that will lead you to have more influence. I would also say the better you are, it's pretty clear that, like when your skill set is stronger, you will naturally have more influence because people will see you as an expert. So tip there is actually to not necessarily focus that much on influencing people, because I think that focusing too much on influencing people is can be a little bit kind of creep towards kind of more like if you're influencing people without the skills to back it up, if that starts to become a little bit more manipulative. But then if you are able if you are really good at what you do, if you are an expert in a specific area, you will naturally start to have more influence. [00:56:00] And I think that it's much easier to influence when you're in that position than to try to figure out some kind of social social tricks to get what you want. So that's what I would say there. And then in terms of leadership, I think that the one thing I've been learning is that there is a difference between kind of heads down writing code, building things and actually the process of managing and making sure people are in the right spot.

Jeff: [00:56:27] So I actually think this is kind of where a lot of like having empathy for other people and understanding people's idiosyncrasies are very important just to make sure that, hey, I know that this person works a certain way. So I am going to interact with them in this in this other way, I think. And then, yeah, like if one person is not really good at being told what to do, you go hands off, you give them more freedom. If somebody needs to be told what to do, then you can kind of get more hands on and kind of map it out for them. And then I think actually kind of actually now that I'm talking about, this is a lot more to say. Kind of like another thing I've utilized for myself is actually asking really good questions. And I think that versus telling somebody what to do, I actually try to ask them to make sure that they come to that conclusion themselves or like I can ask them, ask a question, maybe I'm actually wrong and they can give me some insight that helps me adjust how I think about things. So I found that actually being very pointed and asking great questions is really good for helping lead other people's train of thoughts in the right direction.

Harpreet: [00:57:32] The point about asking great questions, I think is super, super underrated. I don't know what it is, but I've got a bunch of Manti's. You've got like 20, 600 students, and they don't like asking questions. You don't know me like they they feel afraid to ask questions of their superiors or whatever their colleagues, people that they work with. And I think it's that little bit of that imposter syndrome that's kicking him. They're thinking like, oh, man, if I ask a question, [00:58:00] they're going to think, I don't know, something like that's a bad thing. But, yeah, definitely, you know, different kind of question than what you're talking about. But still like asking questions or it's key. You got you have to you have to be comfortable asking questions.

Jeff: [00:58:12] Yeah. Yeah. And I would actually say that, like, it's usually if you're like debating whether to ask a question or not, it's usually better to ask it. And I think that that's one. I would also say, too, is if you're afraid of asking the question, try to answer it yourself through Google. And if you really can't answer it yourself, then ask somebody the question. I think there is a bit of a tradeoff where you also want to make sure I get a lot of messages on LinkedIn from people who are asking me questions that you can probably Google the answers to. So I think there is some nuance there. But yeah, I would always say if debating whether to ask a question, it's usually better to ask it and get the answer there.

Harpreet: [00:58:52] What are some harsh truths about being a data scientist that you want to leave our audience with harsh truths?

Jeff: [00:59:00] Harsh. That's the thing about this harsh truth. So the first one that comes to mind is I actually think modeling is not as essential as it seems and it is Data science space. I actually think that for many companies, many companies can survive without any machine learning modeling. I do think experimentation is extremely important because otherwise you won't know what you want to build. I think modeling is important depending on the company. Like I think Spotify AIs competitive advantage is recommendation's. So that's where like machine learning is the competitive advantage. But I actually think a lot of cases modeling is not really needed. And I know that most people do prefer to do modeling and most of the jobs are not are more analytics like experimentation, focus. So the reality is that there's more experimentation, jobs and analytics jobs than there is machine learning jobs. I would say that is one harsh truth. I think that the second harsh truth is that most [01:00:00] people will have the same background, like most people will have a similar background to you or have stronger qualifications. And the industry is getting more and more saturated with more and more programs producing data scientists. So that is a little bit tougher to swallow, especially compared to when I joined at your data scientist. It was a little bit different back then. So that actually kind of ties back to my original point on find. It's more important to find ways to stand out today like ever before. And I would say the third thing is for more on recruiting, job recruiting is a game that you have to play. It definitely kind of sucks. You have to play it some like you might not like the game, but it is see it as a game that you have to play and you have to figure out how to progress across the different levels and definitely figure out a bowl for everybody who is having trouble finding a job.

Harpreet: [01:00:57] Right. Thank you so much. Last question before we get to a random round, it's one hundred years in the future. What do you want to be remembered for?

Jeff: [01:01:06] One hundred years in the future? What do I want to be remembered for? Great question on just like legacy. So, yeah, I think the first thing that comes to mind is that one hundred years into the future, like if we think back over the last one hundred years, most people who are very successful are not really remembered, especially three hundred years back. It's usually only like really that like the top philosopher, the top teacher. So in a way, I am very comfortable having my name not be remembered in, say, one hundred or two hundred years. I'm at peace with that. However, that doesn't necessarily answer your question. I know where you're at this question. Is it going to get at like what kind of legacy am I looking to really leave behind? And to be honest, I think that's something I'm still like I'm still trying to clarify for myself. I think that the one thing I do do [01:02:00] have on my mind is kind of this idea of challenging ourselves to do things that we don't think are possible. So this is even like not just with career, but anything just to continuously to push it.

Jeff: [01:02:12] Like, I guess I'd want to be remembered for somebody who took on very difficult, ambitious challenges where I knew that I had little possibility of succeeding, but I still did it anyways. And that kind of inspired another generation to really kind of challenge themselves and like, you know. Themselves in difficult situations, so they can see themselves grow. So that's like one that comes to mind. The second one is outside of learning. I feel like learning and teaching is like the ultimate superpower and it's the ultimate positive sum game that we do have. Like when you learn something new and you teach it to somebody else, you learn it better and somebody else actually gets your knowledge. And I actually feel like that education is actually the root of a lot of our problems. If we can make progress on education, we can make progress on all the other problems. So something along that space, I'm not super sure exactly what that looks like yet, but that's another thing that I that comes to mind for me.

Harpreet: [01:03:13] I absolutely love them and it's that's beautiful. Thank you so much for that. Let's jump into the random round question one. Can you give us a rhyme real quick, right off the top of the dome.

Jeff: [01:03:23] I'm doing this off the top of the dome. You see me? I'm Jeff Lee. Harpreet wearing that black sweater. You know what says coffee. You see me right here hitting on that line. You look at me, you know, and not be hitting on the rhyme. You see, I see that bird in the sky I'm flying. Oh, machine learning models. Data science. Oh, you see, I hit another rap session. Oh I got a good ah. Emcee linear regression.

Harpreet: [01:03:53] I love it. I love it. It's got, it's got that proper East Bay, Northern California swamp and [01:04:00] you know thank you, I appreciate it. If you were to write a fiction novel, what would it be about. What would you title.

Jeff: [01:04:06] It turned out a fiction novel. I write about a person doing something extremely difficult and overcoming maybe something like something along the lines of somebody who survived a crazy experience. And I would yeah, I don't know what I would title it, but it would be like somebody surviving in the forest for like two years, living on nothing else, or somebody surviving at sea for like three years just on a boat. I probably write something like something along along those lines.

Harpreet: [01:04:37] It's interesting, man. Thank you. Yeah. What do you think the first video to hit one trillion views on YouTube will be about? And when do you think that will happen?

Jeff: [01:04:48] I probably say I have. I don't even know what the max is right now, but I would probably say like some music video that is top ten on the charts, like top song on the Billboard Top one hundred and some music music video that would hit one trillion. And yeah, it's hard for me to say I'm not sure where it's at right now, assuming it's like it's

Harpreet: [01:05:10] A baby shark and the shark has, I think like eight billion views or something like that.

Jeff: [01:05:17] Ok, and then one was baby shark created.

Harpreet: [01:05:19] Baby Shark was twenty sixteen

Jeff: [01:05:22] Thousand sixty eight billion and about three years. So it's probably like on it. I think it's probably more of an exponential curve. So probably a trillion is still quite a quite a bit of ways to go. But maybe say like six, seven years then.

Harpreet: [01:05:37] All right. I'm actually collecting all these responses and I'm trying to see if the wisdom of the crowd holds true. So this next one is Peter Thiel question. What do you believe that other people think is crazy?

Jeff: [01:05:52] I was wondering if I get asked this on a podcast. What do I believe that other people think is crazy [01:06:00] or very important truth? I believe that very few people agree on. OK, yeah, actually I want to that comes to mind. So Amazon has this like kind of this philosophy of the bar razor where every person you want to hire kind of raises the bar of everybody else. I agree to it to an extent. However, I think it's I actually think I don't know if it's the right frame of thinking when thinking about teams, mainly because every person that joins a team develop synergies with the rest of the team. And if a person can be a bar raiser in one specific role and not a bar raiser in another role, depending on how that role really takes advantage of their core strength. So, like, I think, yes, we should continue to keep making it like raising the bar or whatever, but I'm not sure I necessarily agree with the premise that, hey, this person is a bar raiser or this person is not a bar raiser. It really is contextual and depends on the situation.

Harpreet: [01:06:58] I like that man. Thank you. So what are you currently reading?

Jeff: [01:07:02] So I'm currently reading this book called A Million Miles and a Thousand Years. I read about a guy who is trying to edit his life as if he was writing a story. It's actually learn a lot about just the structures of stories from reading this and wanting to apply some of the principles, such as like every great story a character is like kind of thrown into a situation where you have to experience some type of conflict. And usually the more difficult the challenge, the more that life is at risk, the better the story. So I'm trying to think about how can I apply this idea to my own life to kind of write a better story for myself.

Harpreet: [01:07:41] Definitely going to get that one right now. I just type that into a time. I'll pick that one up and sounds really, really interesting. And what song do you have on repeat?

Jeff: [01:07:51] Repeat. So I was listening to like a deep house mix as I was working today. That's the last thing. Yeah. I don't know if that's a repeat, [01:08:00] though. I would say that other songs that I've been listening to, I've been listening to a best part by Daniel SESAR that's been on a pretty big song.

Harpreet: [01:08:09] I'll definitely check that one out as well. So we're going to open up a random question generator and we'll do a few out of that. So let's go ahead and go to the random question generator. All right. So first question, pizza or tacos? Tacos, mountains or ocean? Ocean. What are you, a natural at?

Jeff: [01:08:32] A natural at maybe like interpersonal skills.

Harpreet: [01:08:37] What incredibly strong opinion do you have that is completely unimportant in the grand scheme of things?

Jeff: [01:08:46] I definitely have an answer to this. I just can't think of it. What incredibly strong opinion do I have that is unimportant in the grand scheme of things? I think I really don't like eggplants. I plants are really gross.

Harpreet: [01:09:01] So, Jeff, how can people connect with you or can they find you online?

Jeff: [01:09:05] Yeah, you can find me on Twitter, I think, on my Twitter dot com as Jeff Lee So m as in Mary. That's where that's what the letter starts with. And you can also check out my website. Cheverly Chronicles dot com and Leha spelled Elai, Jeff Lee Chronicles dot com and I usually will post articles. Any thoughts there?

Harpreet: [01:09:25] Jeff, thank you so much for taking time out of your schedule to come on to the show today. It was an absolute pleasure to finally get you here, man. Thank you so much.

Jeff: [01:09:34] Thanks so much. Have a great rest of the day.