brandon-quach-04-17-2020.mp3

Brandon Quach: [00:00:00] I know that's when the struggle or the mental difficulty happens, when things go wrong. I just know that's not something that I'm going to sit and pay a lot of attention to. And that's just something I'm going to recognize that's there. And then I'm just going to move on.

Harpreet Sahota: [00:00:30] What's up, everyone? Welcome to another episode of the Artists of Data Science. Be sure to follow the show on Instagram @TheArtistsOfDataScience and on Twitter @ArtistsofData. I'll be sharing awesome tips and wisdom on data science as well as clips from the show. Join the free open mastermind Slack channel by going to bitly.com/artistsofdatascience. I'll keep you updated on biweekly open office hours. I'll be hosting for the community. I'm your host, Harpreet Sahota, let's ride this beat out into another awesome episode.

Harpreet Sahota: [00:01:16] Our guest today has over a decade of industry experience as a data scientist and a passionate learner who thrives in multidisciplinary team environments.

Harpreet Sahota: [00:01:24] He's a master of his craft who codes and owns machine learning algorithms from conception to production. He's earned a bachelor's degree in bioengineering from the University of California at Berkeley, a master's in engineering from the University of California, San Diego and a PHD in bioengineering from Caltech. He's worked at companies such as Agilent Technologies, Opera Solutions, Lytx and Teradata. During his tenure at Opera Solutions as an analytics manager, he built sentiment models, threat models and worked in the domain of various online threat groups, including traditional and cyber terrorists. Large corporations focusing on threat analysis for security and business ecosystems and use analysis analytics. He worked on algorithms that detected vehicles, determined a risky vehicle maneuvers and detected risky driving behaviors in cabs. His continuous contribution for celebrated through the achievement of the lyrics go big award for developing a data visualization that enabled a clear communication of complex data analysis to the internal team as well as current and prospective customers. Atara Data. He's currently a principal data scientist and manager, leading the charge to modernize the customer experience by applying machine learning to customer support. So please help me welcoming our guest today. A man who wears many hats, a photographer, a singer, a scuba diver and a data scientist, Dr Brandon Quach. Brandon, thank you so much for taking time out of schedule to be here today. That I really, really appreciate it.

Brandon Quach: [00:02:45] Thank you so much. And thanks so much for the introduction there. It sounds so fancy, but, you know, I was doing these things so early on that it wasn't nearly what you would do if you were to do them today. You write like threat analysis back then. It was just simple keywords. It was simple Bayesian. Like if I see this word more than 50 times, then maybe that according to the data, that would lead to maybe a threatening situation. Things like that. And the field has moved on so far since then. But it's interesting that it sounds so fancy when you say it now. But back then, working on it, it was just us trying to figure out like, oh, we got this data. What are we supposed to be doing with this and starting out with the basics? I just want to say that the things that I express on this show are just my personal opinions. And this is just an interview happening between me and this program. And it doesn't represent any organization that I currently employed by or any organizations that I previously was employed by or any groups for that matter. So, you know, I just wanted to put that out there. This is just me talking, right? This is a data scientist to data scientist.

Harpreet Sahota: [00:03:52] Perfect man, for sure. So you got into it a little bit. But talk to us about how you first heard of data science. What drew you to the field and maybe some of the struggles and challenges that you faced while you're breaking into the field?

Brandon Quach: [00:04:06] Yeah. I mean, for me, there wasn't really much of a moment of breaking into it because there wasn't too much of a field when I started. Right. So I did my PhD at Caltech and I was very experimental. We were taking silicon wafers and we were trying to make them do things like duplicate DNA, using micro-fabrication, things like this throughout the process. I was thinking there were a lot of setbacks that happened in the laboratory setting that wasn't didn't really reflect what I thought I was capable of intellectually. Right. So, you know, maybe I would design a cool experiment, but by the time I did it, then something random would happen.

Brandon Quach: [00:04:43] Maybe I turned on the nitrogen tank. There's no more nitrogen, and everything has been thawed, everything's been prepped, and now it's like you've just lost a whole day's worth of work. And sometimes worse, sometimes weeks or maybe even months worth of work because of that one moment when everything ready to go. Something happened that totally contaminated the experiment. And what things like that happen - and I kind of noticed well, you know, I think I'm more of a like a thinking person. I think I enjoy doing the math more than the experimental stuff.

Brandon Quach: [00:05:10] So I started to look into sort of alternate careers that weren't so experimental. And, you know, consulting was there. There were the legal path was there. There were legal firms would come in and say, well, maybe if you work with us reviewing patents and such for a few years, then you might go to law school after that and and become like a patent lawyer. And so then I was interested in all those all of those things. And eventually I went into consulting. So it was with my first employer Opera solutions now known as ElectrifAi. And we just did consulting for a bunch of different companies. There are a lot of different fields that you had mentioned before, and the title was called Senior Associate. That was just it. And it wasn't really called - I mean, even when I left...it was still called something like analytics manager. It wasn't a data scientist. We never really called it that until probably that transition of when I left and went to Lytx when the title became data scientist.

Harpreet Sahota: [00:06:13] So what were, you know. Along the way, obtaining all these awesome academic credentials that you've had, some of the the struggles and challenges he faced during that time?

Brandon Quach: [00:06:24] Yeah. Yeah. So there's a separate out my academic life and my sort of working life, although they don't really separate all that much. But because in my mind, I think of them as separate environments and doing separate things. So a lot of the difficulties in my academic life have to deal with the fact that, you know, you make this commitment when you do the PhD, it's going to be like around five and maybe even seven years. And along the way, sometimes you wonder. It's a long time, right? You wonder, is this the right path for me? I make the right decision. Should I. Should I just go to work now? Should I get a master's degree and just just go? Should I continue? And if I continue, how long is this going to take and when am I going to graduate? And when what is that breakthrough coming? Because you need something like a breakthrough, either minor or major, for you to be able to publish and for you be able to present to the to defend your thesis. Right. So. During this time period, there was a lot of I'm sort of stuck in and I'm and I'm trying to find my way out. And what can I do to get the good results? And a lot of it, you know, with any science, it's you don't have that much control over things that you do it.

Brandon Quach: [00:07:38] It was all came out. And that's not what I expected you something else, as I would expect either. Right. You do something. I know this is gonna work. You do it. That's not what I expected either. And it just goes on like this for your many years. And that's that's kind of the PhD process. That's that's how you know, that when someone gets a PhD like they, they must have gone through a similar experience to where it was years long. It was you thought you knew everything coming in. Turns out you didn't. You learned a lot about what you can and cannot do. You learn a lot about perseverance and all throughout. You know, if you can make it out and have a nice sort of story to wrap everything around, then you've you've done something sort of significant in terms of like human endeavor. So that's what I like about that working wise. A lot of the a lot of the difficulties are. I would say career related a lot of times as a data scientist, you're working, you're doing things that you're not quite sure if this is real data science.

Brandon Quach: [00:08:38] You're asked to provide some ROI models and say, well, how does that translate to dollars? How does that translate to time saved. How does that translate to new customers that we have got or retention of current customers? And OK, you do that and it's really important to the business. And after a while, though, you start to think, well, how come I'm not building any models? How come I'm not tuning any hyper parameters? How come I'm not doing any deep learning? Why aren't I thinking about whether or not I have the right number of leaves per tree in my random forest and things like this? And that's always a struggle. And that's not a big struggle, but it's something that I think about as sort of a difficulty and thinking How do I do this thing, which I think is what the future employers or what my career is asking me to do, which is build these models? And how do I balance that with what I need today from the directors and the V.P. that say, well, how much money are you really saving us or how much money are you generating for us? How much fraud are you preventing from us? And it depends, right. You know, you could be you could be joining a project or a work.

Brandon Quach: [00:09:48] team or an industry that's already established that those metrics are all worked out for you. You just got to go in, re-train a model with the new data and provide the analysis and, you know, maybe things are a little bit more straightforward. Sounds like to me, right? Things are never as simple as they seem. For me, I've always been attracted to these newer fields going into a company where they hadn't done any data science in this field. And so they have this data set and they think that it's - you walk in, the data scientists are going to walk in and build a model. But what happens is that data scientist comes in and says "Well, what about this field, what does that mean?" And they don't know. And you go around searching for people who know. "And what about this? Why is it this value, not this other value" Then it turns out, well  the data's a little bit dirty. It turns out that people - ideally, you would write you would document things in the right way, but they didn't quite documents it that way. Not everybody is documenting things consistently. So now all your labels across all of the human actions that have occurred are not always consistent with each other.

Brandon Quach: [00:10:53] So you have little difficulties like that.

Harpreet Sahota: [00:10:55] I was laughing at that point about what field does this mean? We don't know. Why is it like this? Yeah. We don't know.

Harpreet Sahota: [00:11:01] It's quite challenging working in environments where there's not an established data dictionary for you to actually know what you're working with. Right.

Harpreet Sahota: [00:11:16] Are you an aspiring data scientist struggling to break into the field, well then check out dsdj.co/artists to reserve your spot for a free informational webinar on how you can break into the field. That gonna be be filled with amazing tips that are specifically designed to help you land your first job. Check it out, dsdj.co/artists.

Harpreet Sahota: [00:11:41] So you mentioned this cool new stuff that you want to be working on. Where do you see the field headed in the next, say, two to five years?

Brandon Quach: [00:11:50] Yeah, you know, I thought about that a lot. I'm not really sure myself, because two to five years ago, everyone was saying that this automation was going to come in, that this building the models, the models were gonna build themselves, they're going to tune themselves and all this.

Brandon Quach: [00:12:06] And I was and it makes sense to me. A lot of the things that I was doing did seem pretty simple. You would do it a couple of models you would choose, you would optimize in this way. You would choose this. You would do this parameter searches and. Yeah. You could. I mean, I was automating them to write. I was writing scripts that would automate all those kinds of stuff. And I thought, yeah, this is probably got legs and this is gonna happen. And so for the last two to three years, I was thinking the next to two years was gonna be about automation and that the data scientists would be akin to a modern, let's say, mechanical engineering who might have done studies in how to - like in fluid dynamics, right. And how to model fluids and what's the pressure and velocity at every point along this wing. But they have software for that. You do do the simulation and you're like, well, now I've got the software. So you're thinking, does that mean I don't need the engineer because the software did it automatically? We're - I think we're gonna get to that phase. But the strange thing to me is that I've gotten the impression that that phase is coming very quickly.

Brandon Quach: [00:13:02] For the last couple of years. So now here I am, right. Fast forward two years and it's not and it's kind of here. I've seen it here and there, but I'm at least I'm still using Python. I'm still coding things myself. And so I think that's what's going to happen, continue to happen in the next few years, two to five years. That being said, you know, maybe the time will come when a lot of these efforts to automate things come into play. And, you know, as I mentioned before, it depends on the industry and the problem you're working on in my career path, since I've always been working on new problems. I don't see it impacting us much, but I'm trying it right. And some of the employers that I'm working for, they themselves are trying to build and have built these kinds of things. And I'm I'm trying to use them as well. And I'm providing feedback on the features that have been developed. And what does it mean to work on a like a real data science project where our ROIs expected it to happen and not just kind of a research thing?

Harpreet Sahota: [00:13:58] Can we automate away human judgment and human creativity for problem solving? Do you see that ever happening?

Brandon Quach: [00:14:05] I don't think that's ever gonna happen. I think the issue is more of people who are thinking of automating things if they're not practicing data scientists and haven't been building models on the ground level.

Brandon Quach: [00:14:17] They may not know about all of the human reasoning and intuition that goes into data science, right. They may - if you ask a data scientist well, what did you do, they might just describe it very mechanically. First, I took the data and I did some cleaning. From there, I chose which variables might work the best. And I have a little algorithm for that. And I got this model and I tried a couple of different models and then we combined the models together. Or maybe it's a machine vision thing, took the images and then I had to do some filtering on the images. Maybe I did some noise reduction. Maybe I kept it raw. Maybe I shifted it into a different vector space. Maybe I kept in the original, maybe we did some sampling and some resizing. That's not all very automatable, right? What I just said, Well, yeah I can write a program for that. Because that's how we describe it. And so then other people would think, oh, yeah, that's it. That's I can write a program for that. But when you really are working on it day to day, all your decisions are in my judgment and my gut feeling, should I do A or should I do B or the model's not performing well. And it could be 10 reasons or 20 reasons. Which reason should I do first? And I only have time to look into two? So it's a lot of decisions like that. And then you got to think about. Okay, I see some result here. I looked into this reason. Now, how do I explain what has happened here? In a way that's consistent with what we've been saying about this system all throughout the project.

Brandon Quach: [00:15:43] So there is that interpretation piece. 

Harpreet Sahota: [00:15:45] As we move towards this new future where a lot of the stuff can be and most likely will be automatable. What do you think is in a separate the great data scientists from just the good ones?

Brandon Quach: [00:15:56] Yeah, I think it's just the ability to think through problems. Most of the stuff that we're doing is just thinking through problems, thinking through "I observe...

Brandon Quach: [00:16:05] I observe that things are not going the way I expected" and thinking about why that is and what's the next best steps for that. And not just from a science perspective, but from a business perspective as well. It could be the next best step is to talk to the business or the practitioners and say maybe the way you're entering the data, we should enter it differently, or maybe maybe the way we're collecting the data, we should be collecting it differently. Or maybe the solution is on the science side. We should build the model differently. Maybe we need some new parameters or we need a new structure in architecture and how we're even building models. Maybe it's on the engineering side. Maybe if we were to speed up the scoring process by a tiny bit, it would make a difference due to some other business factors and how the business speed works. So just that intuition of thinking about what is the next step here. And a lot of that is stuff that you would learn academically, especially when doing a PhD, right.

Brandon Quach: [00:17:01] That's why a lot of data scientists might have a PhD, because during that process is when you're thinking about a years long project - What is the next best move here? And playing that sort of reasoning game. So I think that's what going to separate them right. If you came in and you just said, you know, I learn all the algorithms and I learn how to code things up. I know Python. I know how to code things up. I know tensorflow. And I've done some few projects, but you haven't really done a lot of this reasoning about what do we do now? Now that the results don't look right. What do we do? And that's not something that's often taught in schools anyways, right? In school, the problem works. You just have to figure out how to get it to work. But once you do that now it works. There's not this curved ball of whoops, turns out the data I gave you wasn't really that good. We hadn't been collecting it well for the last few years, although when we hired you, we kind of thought we did. And now do something about that.

Harpreet Sahota: [00:17:49] So it sounds to me kind of like the data scientists that are able to comfortably navigate the maybe are the ones that are really going to be indispensable to their organizations.

Brandon Quach: [00:18:00] Right. Right.

Harpreet Sahota: [00:18:01] Speaking of being indispensable to your organization. Wonder if you could talk to me about what it means to be an intrapreneur and how a data scientist can cultivate the qualities of intrapreneur within themselves and be one for their organization.

Brandon Quach: [00:18:16] Yeah, that's an interesting question. When I think of intrapreneurialism as you're in a company and instead of saying, well, this is the way things have been done and my job is just to rebuild the model and provide it the way things that they've done. And in some environments that's the right thing to do. An intrapreneur is somebody who comes in to a new business problem that says we hadn't used data like this before. And we want to do something with it, but we don't know what. And you take on the mindset of, well, you kind of own this problem in a sense that if this thing goes well, everybody's going to know who are the original people who worked on this. And that would be good for their career right within the company and also outside of company. An intrapreneur is somebody who thinks I'm willing to do whatever it takes. I'm supposed to be a data scientist. But here I need to do a lot of software engineering, here I need to do a lot of communication with the business people, or here I need to do a lot of the operations - analytics ops, dev ops. I need to be monitoring the models. And when they don't look right, I need to be addressing that somehow, and maybe automating that, and getting everyone to work together.

Brandon Quach: [00:19:26] And, you know, whatever it takes to to get the MVP because we're just starting now. Right. So that's the minimum viable product, whatever it is to get the MVP out there so that we can show to the stakeholders the people paying the bills. Right. This is the value that we're that we're bringing with the data. This isn't some notebook that I'm showing you. This is something that's in production and it's making things in real time. And to get to that point is is difficult because there's a lot of sort of dirty work that you have to do to start something up when there's nothing there. There's no infrastructure. There's no team culture. There's no team. You gotts hire the team. There's, you know, very, very little to start with. But it's your vision and your understanding and your feeling of ownership that that drives you to be kind of intrapreneur. And then you can also get your inspiration from intrapreneurs and entrepreneurs. Right. I know that. I like listening to this on YouTube. I Stanford classes on how to start it, how to do a startup listening to entrepreneurs and how they started. They often give you little tips about thin slices.

Brandon Quach: [00:20:31] Little tips about cornering markets where you might get a monopoly and you can bring that kind of a mindset, whether you're an entrepreneur that applies directly right. Now if you're intrapreneur, you can bring that mindset as well, saying, well, well, I'm doing this project within this company. But you know what? Three years from now, there's going to be a vendor out there that says, hey, I'm doing the same thing, or there's gonna be another group within the company, another department that says, hey, what? We're doing the same thing. Well, then how do you want to defend yourself against that internal competition as you go forward in the project as well? So it's just that whole mindset of I'm not just an employee, I'm not just somebody who's doing this small task. I have I feel ownership not that doesn't mean you do everything. That does not what that means. Right. That means that you feel ownership. That means that you discuss these things with your partners, if you will, or co-workers. Right. So I work with a business owner. I talk to that person about it. I was a software engineer. I talked to that person about it. I work with management executives and they represent sort of like investors. Right. And I talk to them about these things.

Harpreet Sahota: [00:21:36] Have you read the book Linchpin by Seth Godin? I haven't read that one now. Check it out, man. I think you'll enjoy it.

Brandon Quach: [00:21:43] Linchpins?

Harpreet Sahota: [00:21:43] Yeah. Yeah. He talks about talks a lot about intrapreneurship in that book. Just kind of the way you describing everything. I think it would really resonate with you. So you had a chance to think about it. It's a bit it's a bit older.

Harpreet Sahota: [00:21:56] It's a bit older, but it's a good book. I was on your blog researching you and I came across your leadership philosophy, which I absolutely loved. I was wondering if you could summarize that for our leaders. I'm sorry for our listeners. Sorry - if you can summarize your leadership philosophy for our listeners and then maybe talk a little bit about how does trust play out in a relationship on a team.

Brandon Quach: [00:22:19] Yeah, I mean, trust is everything. And when it comes to leadership and when it comes to any kind of relationship. Right. That's like a fundamental thing. And trust to me comes from your ability to not be scared of the results that come out of your work or anything that you do. Right. If you messed up, you made a mistake. You're not afraid of that. You come up and you say, well, I was I was hoping to do things this way, but I had made this mistake and I'd done things that way. And you very kind of can't not casually, but you're just not afraid of that. You're not hiding anything. You're not sugarcoating anything. You're just saying, well, this is what happened. And you get that just by having a good track record and a good relationship with your audience or you're talking to, you know, maybe your bosses or something like that. Right. You do a good things and then you make one mistake and it's totally understandable. So that's a trust aspect. And you trust other people then other people were reciprocate to you and then you can trust them, too. You've got to trust them. If there's any time when that no longer happening, then then the whole relationship as it has a major flaw when it comes to the rest of the leadership philosophy that I have, it's you know, I'm big on one servant leadership. I'm big on your you're not the sort of boss. Right. You're kind of the shepherd or you're a helper. Somebody comes to you and they don't even need.

Brandon Quach: [00:23:45] They don't even need to listen to you. You might be formally OK a manager or you might not be for me, the manager. You might be a tech lead who doesn't really they don't really report to you, but you're sort of influential in the area. And it's not about how they do things my way. It's about how can I help this person to grow and to produce their best work when you think of it that way. Then you start thinking about, well, since they have a choice, whether they can work with you or if they don't like you, they could just joined another company. Right. They can just here's my two weeks notice. I've got a better offer. You know, most of the times when a data scientist leaves a job, they're gonna get a raise. So, you know, you got to think about there's actually like a financial incentive for the people who report to you to just leave, report to somebody else, some other company, something like that. They'll probably get a raise from that. You've got to think about when they are with you. It's a it's a privilege to privilege for you. It's a privilege for them. We're in this thing together. How can you help invest? How can you give them the tools to grow their career so that they want to follow you? Right. So voluntarily, they want to go with you once in a while, you're gonna have to bring the insistence that says, hey, I've been with you like 99 percent of the time.

Brandon Quach: [00:24:57] But this one time, I need you to do it this other way. I just know this is the right way to do it, right. And you're gonna call upon that one day. But that's not every day. Right. Because you can't if you try that with smart people, like data scientists are people who think independently, like the people who you want to be hiring, people who have their independent judgment. At least for me. I hate being told what to do. I don't like being told yeah do it this way. I'd rather say here's the goal and you figure it out. And I would figure out. And when I do that at night, I feel good about myself and I feel excited about the work. Like, I'm totally gonna do this. I have good feeling this is going to work. You want to give people that feeling, too, so that they can become excited. Now, another thing I talk about in terms of my leadership philosophy is about how people can judge you later on. Right. Because as you go along in your career, then I'm now in the position of people that have formerly looked up to and have an appreciation for what they had done. And I want the same thing from the people who who follow me now or who I lead now. Right. Because right now they may say, oh, this person's got more years of experience or what have you. And that's why I'm listening to this person's advice. But one day they they're going to sort of grow up, so to speak.

Brandon Quach: [00:26:10] And when they're a manager in a future or a leader executive in the future, they're gonna be able to look back. And when that happens, that's not gonna have any impact on me. So professional meet by then, I'll be at a different place maybe. Who knows? Right. But it's just like a karma thing. I don't want them to look back and say, oh, that guy. That person that I reported to way back earlier in my career, I kind of lost touch with that person and maybe not in my life or maybe they are. But, you know, I can judge them now. I can say, well, that that leadership, that mentorship, the guidance that I received, it was either it was a really, really good and helpful in these ways, but maybe it wasn't so good and helpful in other ways. And I would do it differently now that I'm in the same position. And for me, I just I want that even though it doesn't have an immediate effect on me, it's I want five years from now, 10 years from now, anybody who reports to me or anybody who chooses to to listen to my advice and follow my guidance that they would say, I'm really glad that that happened and I'm going to pass it to the next person. Right. And most of that comes with just that's what happened to me. I received good advice and I received good guidance, and I feel it's sort of my job, my duty to pass it on to the next generation, so to speak.

Harpreet Sahota: [00:27:26] Beautiful man. It's absolutely beautiful. Yeah, you want to be a multiplier of people, right? Like you. You want. You want people who are reporting to you are working with you to become more after interacting with you. Right. That's great.

Harpreet Sahota: [00:27:40] Very beautifully put. Do you have any tips for a newbie who joins a team? What can they do to start building trust with their new found team members and leaders?

Brandon Quach: [00:27:49] Yeah. Yeah. It's just I think it's just the. You're a newbie. You're allowed to make mistakes. Now we're talking fresh - Are you saying new to the career or we're seeing just new to the company,

Harpreet Sahota: [00:27:58] Let's say new to the career.

Brandon Quach: [00:28:00] New to the crew, right? Yeah. Yeah. Then you're you're allowed - you've got a lot of liberties - you're allowed to make mistakes. You're allowed to ask all sorts of questions. You just come out with the honesty and don't be afraid that you don't know as much as some of the more senior folks. Because that that's totally expected. You've got to look a little bit of time here to ask any question. You want to make a few mistakes. You go about your work. You do them. You report the findings as they are. You don't try to mislead or anything like that. And people say, OK, this is great. You've done these things well. These are the things you can improve upon. And, you know, in the next iteration, you should try to work on those things that you can improve upon. So the honesty comes from the fact that you've laid it all on the table here. Good. Get better or, you know, in neutral.

Brandon Quach: [00:28:43] And people can see that you've done that. Now, if you come and on the flip side, you say all the work that I had just done. It's perfect. It's great. No problems. And then I agree. And then two weeks later, maybe four weeks later, somebody else is in that part of the code or somebody else is using that or we're making business decisions on that. And then at that time, we find out something wrong and want to go back and say what which has happened there. And did you did you know that there was this bug or did you know that there was overtraining? Did you know the model was weak in this area? And if you say I kind of knew, but I didn't think it was a big deal. If you say things that sound like, oh, you knew and you didn't tell us, it sounds like, you know, maybe you hid it from us and you don't have to be afraid of that at all because that's something that you're allowed to do. You know, early on, especially, those same rules apply later, you know, throughout your whole career, of course, but especially early on, you have that flexibility of some advice.

Harpreet Sahota: [00:29:34] Thank you. Kind of on that same blog post that I was reading about sad. And you touched on this a little bit. Just a few minutes ago here about how great thinkers abhor being told what to do. And I absolutely love that. If you like, as a data scientist, we need to be able to navigate not with this step by step map with step by step directions, right. But more of a compass. And that may or may not be the point that you are making, but I was wondering if you could kind of expound on that concept for us. Well, why is it that that great thinkers abhor being told what to do?

Brandon Quach: [00:30:08] Great thinkers like to figure things out and come to a point that they believe in the solution. Right. Or in the concept, if you are a great thinker, you'll look at the side A you look at counter arguments, side B. You look at this evidence, you look at that counter evidence and you want to come up with some sort of the thesis that puts it all together. And that's a thesis that you believe in and you feel like I can apply this thesis to a lot of different areas. And it's a fundamental truth. Like I've somehow simplified concepts into a few basic principles that I can I can sort of go by every sort of decision that you make is as a result of your thought and in your head there is this consistent model of how things are working. When somebody comes to great thinker and says you do things this way and that and the great thinker. Doesn't believe in it. Now you're asking to do something that they're not really used to, right. Like every decision you make from brushing your teeth in the morning to how you're gonna do your models, you have - you can explain it to yourself. You know, I'm doing this because of this. I'm doing it because of that. And because of that. Because of that, you have this whole chain that goes back to some basic fundamentals.

Brandon Quach: [00:31:20] And that's your life. And that's how you function. If you ask, hey, should I make a decision A or should any decision B you would you would go through the whole process. If somebody says, you know, make decision B, without you having to go through that, then that's a very funny thing. It's like, how can I do something that I don't, like, believe in? And this is the only time in my life when I'm asked to do this and every other time I'm asked to do things that I feel sort of comfortable with, like I've thought this through. I'm ready to defend these actions. If somebody should ask, why did you do this? I have an explanation and I'm excited about that explanation, too, because you're going to ask me a question. I'm so glad you asked me that question, because I've thought about that little detail that you had just asked me about. Whereas if it's something I was told to do, it's a very shallow understanding. Once you ask me about some details, I might - there's going to come to a point. You know, I didn't think about that. I don't know. I was just sort of told to do it this way. You want to know why? Why don't you ask so and so who told me to do it that way?

Brandon Quach: [00:32:16] And that that's a very dissatisfying thing to say for for people who are used to thinking through everything and having having a comfortableness to decisions that are made because. 

Harpreet Sahota: [00:32:25] You kind of...by just accepting the direction that's given to you, you kind of give up some degree of autonomy. Right. And I think it's that autonomy that, you know, autonomy in conjunction with the mastery of a craft that makes what you do enjoyable.

Brandon Quach: [00:32:42] Right.

Brandon Quach: [00:32:42] Exactly. And then that is what you learn when you go through something like a PhD program. Right. Or or if you don't do that, then it's something that you learn through the experience at work where you tried a bunch of things and things didn't work out right. And then you learned how to navigate through this. Right. And that process takes a lot of a lot of work and a lot of hours, whether it's at work or in academia. And you have to go to a lot of those cycles to build up this model and this mentality. You might say this lifestyle, this way of thinking. The idea that I'm going to try something and it may not work. And I want to try this other thing and it may not work. And after five or six tries, maybe 10 tries. I now have a feeling about I understand what this is about now. I know what it is, and now I can do something that eventually will work. And when it does, I oh, I get it. The reason why all of these happened was because of this chain of events that I can now explain and I can now defend.

Harpreet Sahota: [00:33:42] So talk me a little bit about the mindset of future judgment. What does what does that mean to you? And then how can we employ that in our day to day work as data scientists?

Brandon Quach: [00:33:53] Future judgments to me is, I kind of alluded to before, you do something and there are some things that when you do it, you'll know right away. Is it right or wrong? So I'm not talking about that. That's like sort of Real-Time feedback. I'm talking about the decision that you make. That we won't know. This is the right decision until years from now. And earlier I talked about it in terms of leadership. That happens a lot in leadership. You're giving this advice on how to how to advance your career. Right. Right now, I'm doing a podcast and giving people advice, and they don't know if that's the right device until years from now when they look back. And I'm so glad I did explain to you. I'm so glad I built trust early on in my career. It was so important. You want - and that applies, so that's in the leadership aspect. That happens also in your work. Right. You might be building a model. With that model you're doing research on this model. It's got to go into production. Got to go into testing. It's got to go this and that. Depending on the maturity of the project, maybe the maturity of the company. Hey, it might go into production in a month or it might go into real production in a year.

Brandon Quach: [00:34:52] And at that time, you'll be judged on decisions you had made early on. But there's this mindset, right? And to some extent, if you want to take a lazy way out, you can just say, you know, by that time I'll be gone. By that time I'm on another project, I'll be at another company. These people won't be referring to me anymore. Anything can happen. But I don't like to live with that kind of mindset. Right. That's not my philosophy. I want to know that. I want people to look back long after I've gone and say, wow, that that decision that was made early on that nobody had appreciated, that turned out to be really critical down the road at all or in a leadership role realm. That advice that I received. It sounded so trivial when I was listening to it and I kind of took it kind of casually, but it turned out to have deep consequences in the rest of my career. And that's what I like. And I just like that little game of somebody looking back years from now and saying, I don't know where that guy is, but, you know, what he said turned out to be right. And it would have taken years for me to figure that out.

Harpreet Sahota: [00:35:55] I love that man. That's absolutely beautiful. I've worked in other organizations obviously and when you have a writing code at the top, we'll usually put who is authored by whatever. And I always wonder if somebody down the line is looking back like damn, Harpreet Sahota did some awesome work. And this guy, we're still using his code. That kind of feeling, right.

Brandon Quach: [00:36:13] That's true. Yeah.

Harpreet Sahota: [00:36:15] Yeah, right. That's the approach you should take when you when you're doing your work. Like take a real craftsman approach to it, build something that's gonna last even even just a bit of code, right?

Brandon Quach: [00:36:24] Yeah. Yeah. And I've seen it the other way too. Right. You know, the tech community, depending on where you were, you work on where you live. Right. Like, I live in San Diego. I work in San Diego. And the tech community, the data science community here is quite small. So we're spread out across a limited number of companies around here. And once in a while, I'll hear funny stories about one of my peers had moved on to another company where one of our V.P. is used to work. And now they're delving into that code of that V.P. that that person had written years ago. And, you know, we had I had looked up to this V.P. for so long, maybe. And now that I see some comments that people had made of his own code that I'm thinking, oh, wow, I don't want people to look back at my old code like that. I want them to look back and say, yeah, continue following this person, because even his early code was good and was well documented and made logical sense.

Harpreet Sahota: [00:37:19] But I dig it then I dig it. So shifting gears a little bit here.

[00:37:25] How important is agile and scrum methodology for data science teams? And have you noticed to play out for data science teams?

Brandon Quach: [00:37:37] Yeah, you know, I, I think it's important, but I don't I don't know if it's it's this grand shift in like this grand permanent shift and how we're gonna do things going forward. I, I do see it as a new idea and very interesting and and I'm all for it. Before when I first joined I was unsure. I wasn't that skeptical, but I was just unsure of how it is going to work, because normally as a data scientist, well, as a scientist in general, you get a problem when you work on it. You do some research and it's a very long process. And there's not this idea of - you know, I couldn't describe to you what this, I'm  going to do this. And then I want to do that, and then I'm gonna do that - this is going to take this much time. It's going to take time, this is gonna take this much time and then going to get the result in a month or in six months. Which a lot of the principles - not that I don't say the principles because people can argue a lot about what the principles are. But I can say about the mechanics, the way that we're doing things, we're writing these Jira tickets. We're saying it should take this much time, we're estimating work and then at the end of the sprints we're presenting the work that we have done. So I first I was skeptical about that because I was thinking, how could you know beforehand? You haven't even touched the data.

Brandon Quach: [00:38:52] Now we're talking about I'm going to build a model and this amount of time that it's all right. But I've learned that, well you could adjust for it. Well, I've learned, a few... One, is that you're going to just for it, another is that if you want to be part of an integrated team, then the software engineers are definitely doing something similar. And so are you going to say, well, these groups - the engineers are going to do it this way using agile. But the scientists, you can go ahead and do your research on the side, or do you want to say everybody's doing agile together? And that the scientists are going to adjust things as as needed, but are still sort of same part of the same process? And part of the same team, part of the same standards. So that has benefits on its own. So I found that, you know, it's not as bad as I thought and you know, I, I think it works fine for an early stage data science project. Right. Which is what my my most of my career is, is about this early stage thing. We're trying to get MVP out. You're a data scientist. You do the data science, but you may have to do the other things so everything can be broken down into Jira tickets, it works really well when you're trying to do a lot of engineering work.

Brandon Quach: [00:39:52] It may not work out so well when you're doing the science work, but you can make it work out right. So you can write a ticket that says, you know, I'm going to research this aspect of things and I'm still spend five Jira points on it, for example. And at the end of the ticket, you might say, well, I did the research and I'm going to write another three-point ticket to finish, to follow up on this kind of thing. And I think that as long as you understand that, OK, you're doing science, and that's sort of allowed, even though we're, you know, maybe at the engineering side, they're supposed to say, well, I'm done with this work and I can prove it to you. I can show you I can do a demonstration for your demo. I can run the commands for you. And it works just as I said. Right. Or I'm putting this new button on the GUI its cost you five Jira points. I do the work I can show you. Here's a new button when I see the new button. You're right. It's there. And you know, that's it. But as as long as you we can have the flexibility and adjust the agile process to the work that we're doing as scientists. And I think it'll work out fine. And it's it's a to me, it's a totally fine way of doing things. Now, I've also heard even within the same company. Right. Other people have said, oh, in the past we did we did our agile.

Brandon Quach: [00:41:03] And what I did was I had a thirteen point ticket and I just said, I'm researching. That's like that. That's it. That's ticket. And I mean, that's an extreme way to do it. And I would say, well, if that's how you feel, if you've put in the thought about how is the best way to do this and you feel that, you know, that's the best way to do it. You've read the counter arguments about how there might be a better way to do it. And if you still come to that conclusion. I don't want to prescribe what's the right way to do any kind of work. And I'll say if if you've if you've read the philosophies and you feel like this is the best way to do it and let's do it that way. So for me, I don't feel I think the question was, do I feel that it's really important? How important do I feel right now? I feel like it's as long as you've thought about the arguments that are made for doing things in an agile fashion. And if you agree with them, you're doing them and you can back that up, then that's that's fantastic. If you totally don't agree with them, you come up with solid arguments as to why and you're producing results that are appropriate to that phase of the project that you're in, then, you know, that's great, too.

Harpreet Sahota: [00:42:05] What's up, artists? Be sure to join the free, open mastermind slack community by going to bitly.com/artistsofdatascience. It's a great environment for us to talk all things data science, to learn together, to grow together. And I'll also keep you updated on the open biweekly office hours. I'll be hosting for a community. Check out the show on Instagram @theartistsofdatascience. Follow us on Twitter @artistsofdata. Look forward to seeing you all there.

Harpreet Sahota: [00:42:34] Shifting gears yet again. If we could talk a little bit about grit, mindset, drive.

Harpreet Sahota: [00:42:40] What did these qualities mean to you? How do you think other data scientists can cultivate these qualities in themselves?

Brandon Quach: [00:42:47] That's an interesting question. The cultivate part is interesting because I don't know if I ever spent any time cultivating grit. It's just something that I feel like I had but maybe didn't recognize until later on in my life. Just even things like riding up, learning how to ride a bike. OK, so the way I grew up, I, I just learned on my own. I got on a bike and I fell a ton of times. And in younger years I would blame all sorts of things as to why I fell. I would blame this. I would blame that. But I always just kept going. I always thought, like, well, even though the world's against me and the forces that be don't want me to ride his bike, I'm still going to push ahead because I'm stronger than that, too. Right. I mean, I'm even stronger than that. I'm just gonna forge ahead. And eventually I'm gonna get - I didn't even think eventually. I was just keep going until until it was done. Right. So that was always something that I felt in me. And I didn't, to me, that was just a natural way of doing things. I didn't put a label on it and I didn't know that it was called grit. And that would be helpful later on in my career until it actually happened. My philosophy around that is I think if you if in all aspects of your life, you you show grit if you are willing to.

Brandon Quach: [00:43:55] Live with the setbacks, the pain and the work that's required in any endeavor. If you understand that life is not designed to help you, nor is it designed to hurt you, but you will feel hurt a lot and you will have helped a little. And these things will or will happen. And if you just say, well, yeah, that's just a normal way of life. One thing that helped me is at one point early in my career, like, I forgot what it was specifically, but I think it was something like I was training a model and it didn't work out for like some ridiculous reason about, I don't know, maybe maybe the segments of the population that was the most important. Ended up having the dirtiest data. Something like that. I forgot what it was. Right. And I went to the V.P. and I said, this totally sucks. And I know the segment that I need the most has the worst data. And, you know, this is terrible. What what bad luck I'm having. Then he just looked at me and he said, of course, that's what I mean. Of course, like that. This is this is just a chance, isn't it? Of course not. This is this is life is always like that. This is how life is. And his his mindset of like, oh, just expect it to be like this was different for me. I felt I came in as if I was the victim.

[00:45:04] Right. Like, oh, look, I got bad luck. It had it happened to me or he said, no, no, of course I wouldn't expect it to happen any or any other way. And, you know, harks back to earlier mentors that I had who had said the same thing about Murphy's Law and how anything that could go wrong, will go wrong. And that's just always the case. And I think we have that mentality that I'm about to do something that I think is simple. And I expect that everything is going to go wrong. And I know that's when the struggle or the mental difficulty happens, when things go wrong. I just know that's not something that I'm going to sit and pay a lot of attention to. And that's just something I'm going to recognize that's there. And then I'm just gonna move on. So there's a book I read called Search Inside Yourself, [Chade-Meng Tan] I think that was the authors name. And he was a person from from Google. And he went on to be an advocate for meditation, of all things. And he talked about that saying don't feed the monsters, meaning there's going to be bad things that happen and thoughts that are repeating in your head all the time, which is very typical for a personality like I am. And probably a lot of data scientists are where you do something bad and you just replayed at that moment all the time.

Brandon Quach: [00:46:13] And he said, well, if you have monsters and you feed them, they'll keep coming back because you're feeding the monsters. You don't feed them. They may not go away, but they'll be there and they're just there and that's it. It's kind of a - I don't know, it's like a Zen interpretation or I don't I don't know how to describe that. Right. But, you know, they're the monsters. They're there and they're just there. That's it. So I think about that in terms of the difficulties that happen when you do a project. Right. OK. You've got to step back and it totally bothers you and you hate it. And in younger years, I would said, I hate this. I know I've thought about why do I hate it? And I thought about woe is me. Why did that happen to me? Why? Why is life always like this? And who designed this thing? You know why? Why is the word against it? But now I just see a difficult thing happen. And I don't feed the monsters. I don't think about it. I know it's there. I don't try to get rid of it. I don't try to push it away. It's there. I don't feel it. It's like my my my pet. Right. It's just we're in this together. Me. It's bad luck that I have some of this. Good luck. We're all in this together and this is all part of it. This is to be expected.

Harpreet Sahota: [00:47:12] I love it, man. It reminds me little bit of that quote I think it was Einstein, The most important decision you can make is whether you live in a friendly universe or a hostile universe.

Harpreet Sahota: [00:47:25] So I wonder if you can share some advice or insight with people who might be dealing with the imposter syndrome. And I think you might have touched on that a little bit about not feeding the monster, but...

Brandon Quach: [00:47:37]  Yeah. Yeah. Oh, man.

Brandon Quach: [00:47:39] Yeah. I think the imposter syndrome is something that you'll you'll have all throughout - it depends, right. Some people feel very confident themselves. And I'm the hot shot. And maybe they are right. Maybe they have a fancy title or they're a part of a successful venture and they've got 10 million dollars and they say, hey, I'm not an impostor. I'm the real thing. But I think most of us are gonna feel like, you know, maybe I'm the imposter. I would say you talk with other people and you find out who else feels this way. You go to meet ups, you find other data scientists in your area. You ask them about their experience and you can measure yourself that way. In some sense, you're being more of a scientist by doing that. You're asking, hey, what do you do day to day? And I'm only spending this much time doing models. Are you spending more time? Cause I get the feeling that I'm spending way - I'm not spending nearly enough time that I need to be spending more time with that. Real data scientists would be building models all day and, then they might they might surprise you, and say no, no I'm doing the same thing you are.

Brandon Quach: [00:48:38] I'm doing these reports and finding this ROI and B.I stuff and finding averages and responding, putting out fires. Then you might feel that, well you're doing the same things I am. That's that sounds cool. And every time you find the real deal and I always find them and I always say that person. I think that person's Doing the real thing. That's an actual data scientist over there. And then you get to talk to them and you may find out that they're not actually data sciences or you learn that it depends on the field. That's something that I have recently been talking a lot more and paying a lot more attention about. Right. Somebody who might be and that's a machine vision. Now, that's a field that there is tools. There's a you know, maybe a clear objective. You're going to do deep learning. You're probably gonna do convolutional neural networks. How are you gonna do this? Maybe GPUs. And it's all laid out there. And you do that and then, you know, you feel good about yourself. Hey - I'm doing deep learning. I'm doing all this stuff. Again, but if you're in a new project the way I am, then you should just be aware of that.

Brandon Quach: [00:49:34] Like, hey, not everybody, not every industry, not every project within a company is gonna be that long and established thing. And it's OK. Like, it's also a career path. This whole thing about or I take part in new ventures within companies. I'm an intrapreneur. I'm data scientieer. That's a new thing. And that I don't think that's being an impostor. If you're not doing a classical stuff. Right. Credit card fraud is classical. Customer churn is classical. What else might be right? Recommendation engines are classical. And you might think that I'm gonna be doing those things if I'm really a data scientist. Right. So you think you know maybe I'm not sure - if I'm in a company in a project that's doing that, then sure. That's -maybe I'm a real data scientist. But there's also I mean, a company that is doing something totally different and is applying data science in a whole new way. Maybe in my case right now, it's to automate customer service. And how do we do that? Specific to this company, that's different than the way the vendors are doing it, which is broader to all companies.

Harpreet Sahota: [00:50:31] So, you know, a lot of up and coming data scientists, they have in their head that, yeah we're just building models and all this crazy, cool technical stuff all day long. And they tend to focus on that, thinking that that's what really makes, you know, the data scientist a data scientists, which to a certain extent it does, right. But what is it that's really going to separate these up and coming data scientists from the competition in terms of soft skills?

Brandon Quach: [00:50:54] Okay, yeah, that's a great question. So, first of all, I'll say a little bit about the question itself that you had mentioned. People might think that building the models is a data science. And there's this funny gamification that when somebody comes in interview like I'm asking about the algorithm. Right. Knowing full well that the algorithm - I'm spending, you know, not as much time as, say, future engineering or something else, right. But then I ask, because we all feel that you should know about the algorithms if you're a data scientist. So that's some sort of a requirement. So I'm part of the problem, too, right. I asked about that. And I'm sure everybody asked about any data science interview that you come in will ask about that. So that's just a comment on the question and how it's somehow required still that you know a lot about the algorithms and that that would be all you talk about during the interview, but that your real work is really about data cleaning and reporting. OK. But how do you separate out various - you know, when I when I interview people, I just mainly focus on do I feel like this person can think things through? That's a major thing. And what separates this - If you had thought through everything that has to do with the work that you're doing right.

Brandon Quach: [00:51:56] So and many in most interviews, I'll pick out something from their resumé that they've been working on. And if it's a new College grad, then it can be a class project. That's fine. And then I'll just ask, you know, wacky creative questions about about that. So if somebody recently interviewed me and asked and described a project in which they were trying to locate stitchers in a heart valve. So if you have a picture of a heart valve or or I think it was not a heart valve and maybe some of the tissue or some like that, and then there was an operation where you transplant this tissue somewhere else and then you have these stitches. And they were trying to find where the stitches in this image and the way these stitches are. Right. They have to be biocompatible. So they have to act similar to the tissue surrounding them. So you can imagine sometimes they even look like the tissue surrounding them. So it's a challenge to find out. How do you find the stitches? I had said. OK. Now, what if somebody had come in with 100 new images and you wanted to know what is the qual - And they had labeled them, they had drawn little squares around where the stitches are - 

Brandon Quach: [00:52:54] And you just wanted to know what is the quality of their labeling? That's a very practical question. Nothing to do with any textbook data science that you would work on, but it would require you to think about the problem in a different way. And so then you would have to say, OK, how would I even break that down? What kind of experiments could I run? What is truth here? Because usually the label is the truth. But you're asking me about validating a label. What the heck is that? So you have to come up with maybe creative ways and sort of take your best guess where there isn't a prescribed answer. And I do things like that. And I think successful data scientists can think through any kind of problem surrounding data science, not just the core problem. Little funny things like that. Or I mean, I may have gone another direction and I and they said that they had you some neural networks. Sometimes I ask a funny question. This says, what if I took all the notes if your neural network and it just lined them up one after the other. So if you have 100 nodes, I just have 100 nodes, hidden layers, each one node. Well, what happened there? And then after they answer that.

Brandon Quach: [00:53:55] What if I took this under your nose and said one hideen layer, A hundred nodes long? What would happen there? And then I would say, OK, what if I did that? Then I did something crazy and I said, we are just 50 nodes. I'm just gonna connect them to the output. And these other 50 connected the input. And what happens there? Right. Just some wacky things that in this case, not practical. So the first one was practical. The second case was not practical, but it was something that is so far from what you would normally think about. That it forces you to really go back to the fundamentals and say, what is a hidden node and why? Why do we need them? What are they doing? So that I can get to explaining or guessing as to what would what would happen if I did something totally unreasonable, like like connecting them in a funny way. Right. Because when I do that, then I can see how they're thinking to the problem, how they're breaking down the problem and how they're applying fundamentals. Right. Because if they don't know anything about what is s node and what does it do, then you can't even start to even tease apart how to answer this kind of question.

Harpreet Sahota: [00:54:54] It's really, really interesting, man. I like that line in line of questioning because I've got a lot of mentees and they're always asking what should I study next, what I study next, and I'm like if you're ever at a loss of things to study, go back to the basics and make sure you have those down pat because  the fundamentals are the springboard for everything else. So you need to have all that stuff down pat.

Brandon Quach: [00:55:15] Yeah, you can do that.

Brandon Quach: [00:55:17] You can run a bunch of experiments. That's another way you can do things. You run experiments and then you would when things don't come out right, then you've got to go back and do the same process. Things didn't come out right. Why didn't they come out right? And then you have to know all the different parts of the system to guesses as to which way the system did things go wrong. Want to make it so the software engineering lead and manager that I worked with, he had made he had broken down debuggers into two types. He said there's like the the imagination debuggers and there's the like the brute force debuggers. But it's not that there's two separate types, but this is just two separate. If you were to break it down to two categories, you would kind of break it down like this. Right. Some sort of a, you know, perfect sphere universe model. You want to know, do you have an imagination debugger that comes and thinks about I think the problem is in this part of the code, because I can imagine if it was in another part of the code and this would happen, that would happen in other ways. I'm just I step through this and that worked.

Brandon Quach: [00:56:10] So you get you can study that way, too. You can say, OK, I don't know I don't know what to do here, but I'm not going to give up. I've just got to step through every line of code or, you know, and figure it out. And the first time we did it, we step everything. And the second time you did it, you knew a little bit more about the system. I'm pretty sure it's in this part. I'm just gonna step through this part. Right. And you keep doing that. And there's no real shortcut to this. Right. You keep doing that. You put in the work. And eventually when something breaks, you can imagine. I've I've debug this line by line so many times, but it's in this line of code right here. Not not the line, but I bet it's in this area right here. And you go up and then you know what to do. I know what I'm going to do. I'm going to put a print statement I'm gonna put a break point. I'm going to engineer the code so that it's easier to to find out what the it's easier to sort of plug into. Right. So instead of one function that does a bunch of stuff, I have different functions and I can end the code here.

Brandon Quach: [00:57:05] I can go there so I can engineer things differently, too. So you can go about that way. Right. About bringing going back to what should you do. Right. You should learn how to think through code. How can you learn how to think through code. Well, either you have a built in imagination, you can, you can guess and or it's probably and you have gone through a lot of iterations of code and you can understand the process of how to do that. OK. Therefore you can solve issues when they come up. The issues that you face in real time will never be the ones that you've already thought about because lo and behold, you've already thought about those you've coded around all that stuff. Something from left field comes in and you've got new labels that you've never seen before, or you've got a bug where somebody had wired the neural network in some way by accident. And now you kind of know how to debug that, how to think through that. Right. So it turns out that all these wacky things that I asked for in interviews come up in in the realm of work.

Harpreet Sahota: [00:58:08] Last question here before we jump into a lightning round. What's the one thing you want people to learn from your story?

Brandon Quach: [00:58:15] Oh, man, I haven't even told my story.

Brandon Quach: [00:58:21] Well to learn from this, what I've said is up until now. What is that? That's a nice question. I would say just to expect the good and the bad things that happen with you. I think the highlight would be the idea of the monster. So it's you. It's the bad things that happen to you. And it good things happen to you. You're on this one car together and you're going on the ship. Don't kick them out of the car. They'll do nothing. They're just with you. You just go go with them, with you just expect that it's not just you going. And there's external forces there. This is you, the good and the bad. We're all here together and we go through the trip.

Harpreet Sahota: [00:58:56] I love it, man. I love it. Yeah. That concept, that monster. I gotta check out that book, man. So I'm  definitely be looking for that. That really resonated with me. So let's jump into the lightning round here real quick. What's your data science superpower?

Brandon Quach: [00:59:11] Oh, oof. Data Science super power? Willing to do non data science work when needed, but at a reasonable amount of time, if it's the whole thing, then eventually I'm going to have to do some real, real data science work. Really to, once in a while when needed, go and do any kind of other work.

Harpreet Sahota: [00:59:28] What's a topic, academic or otherwise, outside of data science that you think every data scientist should spend some time researching up on outside of data science?

Brandon Quach: [00:59:40] I mean, the obvious answer would be like businesses, or something like that. But if you say data sciences is all of that, then if I really go outside, you know, maybe research in body language. You explain things to people and get a feeling about do they understand what you're saying? Because a lot of times as a data scientist, you're talking to people who may not be data scientists.

Brandon Quach: [00:59:58] They may not. They may. They don't want to say, right. They're OK. Stop what you're saying. I don't understand any of it, but they may not. And maybe if you read a little bit more about body language or not knew more but paid more attention to that. Maybe you would pick up cues and you could stop yourself and say, hey, do you have any questions about what I just said? Right. And maybe go from there.

Harpreet Sahota: [01:00:16] What's the number one book -  Fiction or non-fiction or both - 

Harpreet Sahota: [01:00:21] You would recommend our audience read and your most impactful takeaway from it.

Brandon Quach: [01:00:28] Number one book. There's so many. Which one? Fiction or nonfiction. There is a book that I  - first, can I, do I  have to choose only one. I probably have to choose...

Harpreet Sahota: [01:00:39] No, no you can , you can

Harpreet Sahota: [01:00:42] I'll let you, I'll...

Brandon Quach: [01:00:44] Okay. I'll just use one. So I liked, Case in Point which is a book that I read -  it's an interview book - and it's a book that you would read to get a consulting job like at BCG McKinsey or one of those and it breaks down how to do the case study interviews. And my big takeaway there was the sort of breakdown, any problem into its individual - any business problem - into its individual pieces and attack each of them individually. And it just goes through various cases and says like, OK. So and so company is trying to analyze why subscriptions have gone down for this particular month. And then you look at how how how do I begin to tackle why subscriptions have gone down. So they have that. I think the author was Constance. And you know what I figure with the others, but I can get off line and try to find it. I don't think many people would give you that answer. It's not a lie. It's not a bestseller, I don't think, at all. But it's just something that when I was trying to get into consulting and I had gone do a lot of case study, interview, mock interviews, and I found that book was pretty cool.

Harpreet Sahota: [01:01:52] Definitely man, I'm gonna get that book right after this. So if we can somehow get a magical telephone that allowed you to contact 20 year old Brandon, what would you tell him?

Brandon Quach: [01:02:03] I would say just keep keep doing things that you think are interesting. Just keep doing that. Because 20 year old Brandon was not doing data science. Right? I was doing bioengineering. I was making guesses about the future. Getting all those wrong. I was just studying things that I thought were interesting. I was studying things that I thought were the future, bioengineering, which maybe is the future. But I didn't it wasn't as big as I thought it might be. And in the end, I went into this data science, which at the time, at that time would never have even existed. Right. So I will just say whatever was interesting today, study that. But it's interesting, as a kid, I don't study something that you think is what I was 20. You know, computers and Silicon Valley is already around. So I could have just gone straight to computers at that time and done computer engineering. At that time, though, I didn't sort of feel it in my heart. I didn't. I felt that I wanted to do something, something else. And eventually I know I'm in that field anyways. Right. But at least I got to go through and study what I like to study and gave me a foundation of how to think of the World Foundation mean because I learned all sorts of stuff about how like I don't know how to solve mechanical, like biomechanical problems or there's a sphere sitting on a ramp at 20 degrees.

Brandon Quach: [01:03:13] And at what speed is that sphere going to roll down this ramp? And you know how to do those calculations. And I just even though it wasn't data science, it wasn't computer engineering, I could still use that skill set in the field that I'm doing now. And it also taught me how to enjoy things. Right. Like, I enjoy this. We're going to join that work. But from the jump from the get go, I said I'm I'm in college. And at the time, it's computer science. But today might be data science. Hey, data scientist makes money. I want to do data science. And you do that for a little while and then you say, you know what else makes money. This other thing makes me want to do that. And now all of your decisions are based on what makes money and you're at sort of a disadvantage. Because you're competing with people who could like this stuff and who enjoyed us work. You know, my guess is that those people would do better.

Harpreet Sahota: [01:04:01] Yeah. Because you have to in order to excel in anything that you get to operate out of that place of happiness and joy. That's when the growth in anything will occur. Right. You're not facing the resistance. You're allowing yourself to thrive and blossom.

Harpreet Sahota: [01:04:14] That's great advice. Now, what's the best advice you've ever received?

Brandon Quach: [01:04:19] I mean, one thing that the first one that pops into my head is something I repeat a lot is I went to a VP and I explained here's the whole situation. And he said, you got to give options. Just come to me and say, Option A, here's the pros and cons. Option B, here's the pro's and con's, and here's option C. Which one do you want? And as I thought about that and I think about what I see in the movies, about how somebody might go up to the general and say, hey, general, a lot of stuff is happening in the field. I could do a or I could be. Which one? Where should I go? And a general would say, OK, you should do it. That's it. Then you go back and you execute a I think that's that's something that I often tell people around me and people would you know, it's just a nice way to interact with stakeholders. Right. With business people, especially as a data scientist. And like any scientist, you want to go into details and you want to go into the interesting bits and stuff. But what people really want to know is they want to make a decision. They should it should be A, should it be C, should it be C, should be a combination. And what are the implications for each of those? Right. What are the costs and benefits?

Harpreet Sahota: [01:05:22] That's really, really good advice. What motivates you?

Brandon Quach: [01:05:26] What motivates me. What motivates. Oh, OK. What motivates me is the idea that I'm going to do something that I'm proud of myself. Right. Not something that's I hope somebody else like. It's not something that I think somebody else would like. I do something that I say, hey, that's cool. Even if nobody else notices what this thing is like, I'm so happy about the work that just happened here. And, you know, it all ties together, right? Maybe five years and somebody realizes this is a good decision. But right now, I'm happy with that. And for me, that the nice thing. That's that's what motivates me at the individual level. Now, me as a leader, what motivates me is watching people grow, right. People who choose to follow you, watching them grow. Making sure that that they're getting all the advice they need. And that one day they're gonna look back and say and tell the story of how an early mentor, mentee, or manager, even basically an early leader that they had chosen to follow how that person impacted their life in a positive way. And they might tell their kids right. And their kids don't know me. But I had this thought in my head that, hey, one day they might do that and know that would be cool.

Harpreet Sahota: [01:06:40] I love it, man. Especially the thing about doing something, because you want to feel proud of doing it. I think that's definitely a driving force behind me creating this podcast. Like, I don't care who listens to it. I listen to my own fuckin podcast because there's so many cool guests on here. You know, I learned so much from from it that comes on my show. And it's something that I'm truly proud of making.

Harpreet Sahota: [01:07:00] And so, yeah, that that really resonated with me. What song do you have on repeat right now?

Brandon Quach: [01:07:05] Oh, man, I have a ton of songs. Oh, OK. Funnily enough, we're doing all this social distancing right now from a distance by Bette Midler. I have that, I repeat, because I, I'm trying to sort of keep. So Bette Midler, I think she also plays a role. One of the movies where she's trying to keep the troops entertainer or what have you. So I'm trying to keep my troops entertained. My people entertain. And one idea that I thought of early on was I tried to, like, sing a song to the people at one time. If it gets to that point, that they would need that. So I called up one of the people I worked with and I said, hey, you want to do a song? If you do the guitar, do the vocals. We'll try to do from a distance because, you know, social distancing from a distance, it kind of matches out. Right. And so we have this little side project going on. I have decided if I want to do it yet, because it's a big step to be singing in front of your group. But I figured, you know, if we need it, that that uplifting or even a funny moment, right. Where everybody's laughing at you because it went so terribly wrong or something, if we needed that, I have that in the tank and ready to use its in reserve.

Harpreet Sahota: [01:08:07] I hope this goes viral on LinkedIn.

Brandon Quach: [01:08:11] I hope not.

Harpreet Sahota: [01:08:14] So how could people connect with you? Where can they find you?

Brandon Quach: [01:08:18] LinkedIn is a good place to find me, just Brandon Quach. You should say that with maybe TeraData or Lytx then that would, that would come up.

Harpreet Sahota: [01:08:29] Brandon, thank you. Thank you so, so much for being so generous through time and being on the show today. I think people have learned so much from from everything that you've said today. So thank you. Thank you.

Brandon Quach: [01:08:41] You're very welcome. And I sure hope they do. And if you hear any any good feedback or even any negative feedback, feel free to send out my way. I'd love to hear about that. And, you know, we were barely scratched the surface. You know, I wouldn't get time to talk about a lot of things, but it was enjoyable.

Harpreet Sahota: [01:08:58] A man well, Part two. We'll bring you back on for part two for sure.

Brandon Quach: [01:09:03] Let's do that. I wanted to be. I wanted to be demand driven. Right. So enough people say they want a part two, we'll do a party two. I'm not going to force part two on anybody.

Harpreet Sahota: [01:09:14] We should do the part two just because we want to do it, man.

Brandon Quach: [01:09:17] There you go. Hey, that's that's. Then you've got you put it back on your ready to have it in the archives. And one day when people say I want to see it. What is that? Oh, we already did it so much that we did it anyways. Since you want to see it here it is right on that. Really.

Harpreet Sahota: [01:09:34] Thank you again for your time. Really appreciate it.

Brandon Quach: [01:09:36] No. Yeah. I appreciate it. Awesome.