Speaker 2: (00:01)
You know, I say a lot of things that people hate in the field, so when I say talk in a border plan, sometimes infuriating, but I see the field we're running into a giant train road and that is not what everyone wants to hear. We have a huge problem 

Speaker 3: (00:20)
[inaudible] 

Speaker 3: (00:32)
[inaudible] 

Speaker 1: (00:38)
what's up everyone? Thank you so much for tuning in to the artists of data science podcast. My goal with this podcast is to share the stories and journeys of the thought leaders in data science, the artists who are creating value for our field through the content they're creating, the work they're doing and the positive impact they're having, the film their organizations, industries, society, and the art of data science as a whole. I can't even begin to express how excited I am that you're joining me today. My name is Harpreet Sahota  and I'll be your host as we talk to some of the most amazing people in data science. Today's episode is brought to you by data science dream job. If you're wondering what it takes to break into the field of data science, checkout, DSDJ.Dot.Co/artists with an S or an invitation to a free webinar where we'll give you tips on how to land your first job in data science. 

Speaker 1: (01:29)
I've also got a free open mastermind Slack community called the artists of data science loft that I encourage everyone listening to join. I'll make myself available to you for questions on all things data science and keep you posted on the biweekly open office hours that I'll be hosting our community. Check that out@artofdatascienceloft.slack.com community is super important and I'm hoping that you'll join the community where we can keep each other motivated, keep each other in the loop on what's going on with our own journeys so that we can learn, grow and get better together. Let's ride this beat out into another awesome episode and don't forget to subscribe, follow like love rate and review the show. 

Speaker 3: (02:14)
[inaudible]. 

Speaker 1: (02:29)
Our guest today is an engineer who loves solving problems just as much as he loves sharing his knowledge on applied machine learning and educating business leaders on its value. He's an applied data scientist and 

Speaker 1: (02:39)
machine learning practitioner who has been involved in the world of technology for over two decades, both in hands-on and leadership roles in the last decade. He's been a leader in the space teaching companies how to build machine learning capabilities and integrating those new technologies into their businesses. Having brought products to market with annual revenues in the hundreds of millions of dollars. He's known for his blunt talk on machine learning biases and information security, as well as his musings on how to hire talent, how to run projects, how to break into the field, and how to make machine learning profitable. If you've been published on mainstream platforms such as fast company, Silicon Republic, Katy nuggets and open data science since 2015 you've been recognized as a top voice and expert in data science and machine learning. Having spoken at conferences, business seminars and academia for the better part of the last decade, she served in roles from team lead to senior principal to chief data scientists and as founder of V square machine learning consulting. So please help me in welcoming our guest today. One of LinkedIn's top voices for data science in 2019 then Vashistha then thank you so much for taking time out of your schedule to be here today and I really appreciate you being so generous with your time.
Speaker 2
 I appreciate you having me on. Thank you. 
Speaker 1
So let's go ahead and get into it. I want to first start off by talking about how you first heard of data science and what drew you into the field. 

Speaker 2: (03:58)
You know, it's funny, when I started data science I call myself something else because I didn't know data science was back in about 2011 just getting out of, I was in kind of this transitional phase from being a software development leader, kind of going into the strategy side and the product management side and who was using data on a regular basis to drive strategy, drive decisions, understand customers. And I come in to fill into data science and like I said, I have no idea that there was this rich field out there because it really hadn't gotten a lot of traction. People talking about data science, they're just saying big data. 

Speaker 2: (04:38)
We're still talking about business intelligence sorts of things. And so the way I got into the field was really convincing a lot of companies to hire me to do, uh, you know, this sort of data science thing that didn't have a name yet. Then we're running into others from Google, Facebook, from other large companies that were starting to work on big data, data science, sort of the evolution of finding all of the tools that were out there, working with other groups that we're creating tools. And it's sort of, like I said, it was this evolutionary process where in about 2012, 2013 I found out, Oh, I'm a data science. And so, you know, it really was one of those things. I woke up one morning, a data scientist. 

Speaker 1: (05:22)
That's pretty interesting, man. So having, having started from a time, you know, when data science wasn't really a buzzword or a, you know, a hot phrase or hyped up, where do you see the field headed in the next two to five years? 

Speaker 2: (05:37)
You know, I see a lot of things that people hate in the field. So when I say talk is a borderline, sometimes infuriating, but I see the field running into a giant train room and that is not what everyone wants to hear. We have a huge problem right now around sort of the use of machine learning is analytic and so right now we are using a whole lot of very, very complex math and a whole lot of high end compute to create analytics. These are models which do not generalize well to new data and new circumstances code. It is a great example. You're seeing all sorts of marvels right now integrate because they do not generalize very well to novel situations. They've truly not learned anything. They're simply parroting and you can see the quality models now standing out from those which are really analytics true because the quality models aren't performing perfect. 

Speaker 2: (06:36)
You know, one could have foreseen and no model really has the data to understand in general. And so while we're seeing the limitations of models, we're actually seeing which ones are, and that's why I think that as a field we're on our way to train because we have invested a lot of months saying these machine learning algorithms are robust and they've actually learned some. And in times like this we truly reveal the limitations of machine learning. And we have to come to a sort of record where we say, look, machine learning has practical applications but only if you understand what it is you feel you can use complex math to drive value and in the pure analytics space. But when you branded as something else, when you call it machine learning, you give it this extra sort of responsibility to generalize and to understand and to have actually learned. So I can apply that learning to normal situations, situations. However, we are not doing that very well. And so we are seeing sort of the beginning of this train, where can we separate out complex analytics from true machine learning.

Speaker 1: (07:44)
Wow. That is a very, very contrarian view in a very interesting take on, on what the future is for data science. Thank you for sharing that, but in that vision of the future that you have, what do you think is going to separate great data scientists from merely good ones? 

Speaker 2: (08:02)
You have to understand patients with the algorithms that you'd implemented and the limitations of the data. Everyone makes decisions about both, but a true great data scientists, and I wouldn't even call myself the truly great data scientists. I watch others who do this in ways that I would like to emulate. I won't call anyone out, but there are some truly great companies by themselves that run their own consulting companies. You look within Google, you look within Amazon and you will find some truly amazing data scientists, machine learning data scientists who are conscious of not only what the limitations of their data  but also what the limitations of their specific approach of algorithms and the deep learning models that they've created. They all have limitations and you make decisions consciously about what you build into your models and the great data scientists. What you really want to aspire to and what I aspire to. 

Speaker 2: (09:03)
Becoming someone who is better at understanding the limitations and architecting solutions that minimize sort of the adverse effects of some of those limitations until we've built better models have built more of a, a smarter machine learning, which truly understands sort of the rough edges of causal relationships and so we've built that and we have a sort of smarter framework and approaches that we use, the smartest things that we can do in the most useful things that we can do. Really relate to understanding limitations and crafting solutions so that those limitations have the smallest impact possible. 

Speaker 1: (09:54)
Are you an aspiring data scientist struggling to break into the field or then checkout DSD J. Dot. Co forward slash artists to reserve your spot for a free informational webinar on how you can break into the field that's going to be filled with amazing tips that are specifically designed to help you land your first job. Check it out. DSDJ.Dot.Co.artists

Speaker 1: (10:20)
That's very, very interesting. Yeah. Like I'm understanding limitations and how to work within those limitations I think takes a fair bit of creativity. And I like what you mentioned there, how you are taking from other data scientists out there in the industry kind of taking or learning from them rather. To me, creativity kind of feels like taking different bits from other people and synthesizing it to your own problem statement but still coloring within the lines and limitations of your own problem statement. Um, I was wondering what role does being creative and curious play in being successful as a data scientist and how can someone who doesn't see themselves as creative be creative? 

Speaker 2: (10:58)
In my opinion, creativity and in my experience I should say my opinions are shaped by all of the things that have happened in the last 10 years. Creativity as a result of the failure, repeated failure and admission of those failures. Heatedly found myself over estimating my capabilities for the forest. Maybe three years is my, for my career in data science. 

Speaker 2: (11:23)
And that was the best thing that could have is true. Push the limits of what I could do and then begin honestly assess sort of in a post-mortem way, what is it that I've built, what have I created, what were the limitations of those things I had done and how am I going to get it? You're right. It's a synthesis of not only your own personal experience, what's worked, be very honest about what hasn't worked, but it's also looking at people who doing sort of more than than research for building applied solutions, building the sorts of models that get into production. Okay. Provide value to a business and looking at what worked and what didn't. Why did it take longer than it should have? That's creativity is recognizing the law and being relentless about finding the flaws and finding ways to pick apart what's been done in the past. Skeptical in an almost, you know, being adversary or that has a totally different meaning in our field. Being a adversarial to your own work and to the work of others at the same time, welcoming in new ideas with that same sort of adversarial approach to looking at and say, no, that's wrong. I don't like that. That doesn't seal proven enough. It doesn't feel like it works well enough. I feel like I'm doing too much work to get too little accuracy. I feel like the accuracy isn't providing business value, you know it's attacking the way that you think consistently reforming it to be something that practical and more useful. It's more real tangible. That's our creativity. 

Speaker 1: (13:04)
I love that man. That really resonated with me. I've got like chills going on right now. I was talking to talking to a junior colleague of mine earlier today actually and I was really communicating that same sentiment that you're saying it's, you know what makes a great data scientist is not a PhD. It's not, it's not how many courses you've taken. It's the ability to really like you say, be adversarial step step outside of what you've done and question yourself like, Hey, was this actually right? Am I just seeing something that that is just a fluke? Did I actually do this thing right? Like having that type of of self criticism I think is crucial. Speaking of like PhDs and masters programs and all that, you know most up and coming data scientists tend to focus primarily on the hard technical skills and they think that's what's going to separate them from the rest of the crowd. But what are some soft skills that candidates are missing that are really going to separate them from their competition? 

Speaker 2: (13:56)
Communication but overwhelmingly communication. If you cannot talk to people outside of the machine learning organization, you cannot explain instilled value of machine learning into every facet of the business. If it is singularly focused, if it cannot, you'll get past its use as a cost saber. If it cannot get past its use in a single feature or supporting a specific part of a product, if it cannot become part of a company, it will fail to achieve its maximum impact. And so communication for a data scientist, an individual beyond just being a code monkey, it gets them past sort of that, that pigeonhole of staring at the screen, looking at the data, watching the ticker tape of accuracy and metrics of bottom of screen. You have to go beyond and talk to users and people who don't really understand machine learning and what the use cases for machine learning. It is all about getting outside of the lab, getting into the field without communication. There are so many barriers in front of you to do that. No one wants you to talk to customers if they're scared of what you're calling to say no one wants you to talk to the C suite. 

Speaker 2: (15:23)
Lock the closet. Yeah, it's, it's really, it's funny you see women excelling here because there is so much more of a focus on communication skills, sort of as an advantage builder for everyone trying to get into the field. So as we get more women into those shields, you will see more of this expert communication coming out. They're sort of leading the way. It's interesting. I haven't really figured out what cause behind this as it's still trying to understand it, but it seems like women are leading us in communication across sort of technology field but a machine learning and cybersecurity. Some of the best communicators and some of the people who are doing the best job teaching are women in machine learning, data science and information security. I follow now more women in those fields than I do men and it's interesting how much better a lot of women are at communicating. 

Speaker 2: (16:27)
The men are, like I said, I haven't figured out where that comes from, but it seems like the focus is more, and this is again a great pitch for diversity. It feels like the more diverse the team, the better that communication and a lot of time women are driving that because they understand for some reason better than guys do that there is this need to the skill that's necessary. And again, I can't explain why, but it really does seem like women are doing a better job. I can give specific examples. No. All the way back to, I'm thinking of people who I watched in the field eight years ago who really communicated well enough that I began to understand the field itself better, but I began to understand how I could communicate better with other people outside of this. Like I said, it wasn't interesting sort of anecdotal piece of the field to see that diversity created better communication, put a better emphasis on communication and so if you're talking about soft skills, really look at communication and really look at building a diverse team because a lot of that diversity reeds to better communication not only inside of the team but externally. 

Speaker 1: (17:46)
100% agree with that and communication is so key. I was talking to 

Speaker 1: (17:50)
a colleague of mine, they'd sent me to to an offsite office talking with some stakeholders that were going to be the main consumers of this product I was building out and it really just felt like as soon as they got to know me, understand me and trust me and trust the way I was communicating what I was doing, that they were able to garner more trust. Not only in me but in the product that I was creating for them. So yeah, communication is definitely key. 

Speaker 2: (18:14)
I was thinking of Carla Gentry, I'm sorry, that was the name I forgot earlier on. 

Speaker 1: (18:18)
So I was doing some research and I came across, I think it was a blog post that he had written about the growth mindset. I was wondering if you could talk to us about how you first got introduced to the concept of growth mindset and why it's important to understand for people who are in the job search. 

Speaker 2: (18:32)
It's really a big piece of the strategy is mindset. It's understanding how a company can create different pieces of strategy. One of those pieces of strategy is the mindset of the com and understanding that a mindset has to be, and um, Microsoft did this really well when they branded themselves, they created the challenge of mindset 

Speaker 2: (18:57)
and they decided that they were going to admit truth. They were no longer the incumbents. They were no longer at the top of the heap. And in order to regain dominance, they had to adopt challenge your mindset. Excellent. And so a growth mindset can help in the same way as a strategic practice at companies. But for an individual, a growth mindset sets the tone for their entire career. Are you someone who thinks you continually learn and improve or are you someone who gets to a plateau and says, this is all there is to know, therefore I know it all and that can be a limiting sector in your career to believe that there was always something left to learn. You use, you're continually reading, you're continually learning and improving your skills. You're continually getting better. It's to believe that there isn't enough. Everyone is to believe that four us to win, there is an outcome where you've benefited an added benefit. 

Speaker 2: (19:58)
There is not sort of this zero sum game where in order for me to win, you must lose or someone else. And this is, and again this goes back communication. If you are of a growth mindset, you're not only will want to teach, you want to learn and those two pieces of communication are essential. It will mean that you'll stop talking. And I'm in one of those places in my career where I've learned the importance of shutting up and sort of, it's interesting, I'm learning more than I ever have before. Now that I'm kind of saying, okay, I need to shut up. I need to amplify smarter people than I, and this is growth. This is instead of reaching a plateau and saying, I'm done, you know, I'm coming back to I'm an EMBA, so I need to learn again. I need to do, I need to sort of really fresh my learning and understanding things like diversity have really come to the front of my mind over the last couple of years. 

Speaker 2: (20:58)
And I look back over my career and I said, Whoa, what could have been better if I had work better with more diversity, could have been better. And again, it's back to that whole picking apart every piece of what you're doing and saying, is this really work? And diversity is one of the things that I keep coming back to because I see sort of the poison that it can introduce to datasets. We're not just talking about you know, discrimination against classes or women, but really looking at we excluded something from this data that doesn't necessarily have to be discriminatory because a lot of times when you talk about diversity you'll meet. We come to the opposite of, but I think that sort of this inclusiveness is part of the abundance and part of the growth is beginning to listen to more voices to include more in our datasets. You can see growth and abundance really have so many places. They can help your career and sort of help you pick apart and do whatever. I was talking about what went wrong well could have been better. What would the team have look like? If I had included this person, what would the data set have looked like? I had considered more Chase's had I gathered more data how to took samples different and this is growth. This is continuing to look back and say, I can do better. 

Speaker 1: (22:23)
That's awesome advice.. Beautifully put. I appreciate that. Kind of shifting gears a little bit here. Can you talk to us about how a up and coming data scientists can tie a particular ability or a particular requirements with a business need specifically in in cases where one doesn't have any work experience to speak of? 

Speaker 2: (22:40)
I did a little research on this has been kind of an obsession of mine for four years, no more than four years. Well it's been a long time trying to tie capabilities to business outcome and business strategy to capability because those are really the tie between strategy and execution and sort of that execution piece coming back around and driving the next round of a strategy. What does an individual, how did they play into that whole sort of circle and a lot of understanding how capabilities tie directly to business value is making that connection to the actual, the actual driver, the actual that you are implementing from strategy. And so it takes a lot of understanding and this is where business acumen fits in to the the machine learning and data science skillset. If you don't understand the business, if you don't understand the use case, you don't understand the user, the product, it's a holistic view. Pieces of that, your solution. 

Speaker 2: (23:47)
We always have imperfections very significant come from not understanding how your capabilities and how your models contribute to the bigger picture, to the larger goal of the, and so if you are an upcoming data scientists notes, you are actually in a better place than most data scientists, machine learning engineers because oftentimes again, it's that silo is that lack of external communication. It is the lack of interaction with the outside world. And again, talking about diversity, not just the traditional definition of diversity but diverse datasets, diverse views, sort of getting diverse requirements. If you can, as a aspiring data scientist kind of integrate that into your approach and into what you do every day, you will be better in interviews, you will be better in the business. You create more value because you can understand the bigger, you can understand not only how you are fit, but how you can apply them in order to create business value. And that's a huge leap. 

Speaker 1: (24:58)
That's really interesting. Man. Thank you for that. Um, could you share some tips or words of encouragement for our listeners who've got a couple of decades, let's say 10 to 20 years of traditional, and I'm putting tradition on air to coast here, a traditional ITŐs under the belt who are now trying to break into data science. What challenges do you foresee them facing and how can they overcome some of those challenges they built?

Speaker 2: (25:21)
That has to work. If you're in it right now, if you're in software development, your infrastructure is your end one of those technical engineering roles you've actually had to build. They've actually had to work. And if they didn't, that's that experience right there is massive because so much of what is done, it's hypothetical, it gets into production and it looks nothing like what it was to begin with because it had to be shaped and molded and slammed into place and everything else. And really what I can say is do not worry about the barriers to entry. If you understand linear algebra, amazing, great. You will have a deeper comprehension of the models you put into production. And that's what you need, but you don't have to start there. You have to start by building things. Does that make it into production and work and provide and as a traditional it purse, any role, any technical you've had to deliver and you've gotten, you know especially if you've done this for 10 years, you've had them deliver your whole career and if it didn't work [inaudible] you were working somewhere else the next week you were not there anymore and so that is so relevant to making data science and machine learning really work for an organization but it's often a disconnect. 

Speaker 2: (26:54)
People in my field, me included in some cases have had where you know the shiny object became more important then the dollars that it drove. 

Speaker 1: (27:04)
That's very good insight and I'm sure a lot of people out there are going to find some encouragement from that statement. You mentioned barriers to entry. I was wondering what advice or insight you could share with people breaking into the field who are looking at these job postings? Some that seemingly want the abilities of an entire team wrapped up in one person and they end up feeling dejected or even discouraged from applying. 

Speaker 2: (27:28)
Well, the last posts that I wrote talked about troll jobs and I think a lot of those jobs that you're talking about, the ones where you look at it and go, that's wait, you want 12 years of experience with a technology that's eight years old? Well that's cool. You see that a lot. Those are trolls. Those aren't real jobs. There's nothing behind that job. If you were the most successful person on earth in data science and you decided on a Lark to apply for them to exclude me, I've done this a couple of times. You find out on the other end they'll say something, you know, ridiculous. At the end of the process, uh, you need to work for this much an hour. So below what a normal hourly rate is that you go, okay, you weren't really trying to hire. And so we have to avoid the trolls. 

Speaker 2: (28:15)
And as soon as you start looking at legitimate job posts, you begin to get a whole lot more encouraged by the fact that companies like Google and Facebook are dropping those barriers to entry. And again, especially if you are someone who is underrepresented in data science, if you sit down in an interview and you don't need to at least have some representation, you don't have someone that looks like you or someone who speaks like you or someone who is from your background. If there isn't somebody like that in the interview, just get up because those teams are toxic and we need to, those teams either need to change and change the requirements and changing their thinking and changing the way they hire. I mean if you sit down at a table with six PhDs and you have a master's, just that is not the right table for you. Similar if everyone is hiring themselves over and over again, it's not a team you want to work for and you won't find that. So always look for teams where at least one or two people are like, because those are the teams that have created sort of the diverse environment, not just the traditional sense of diverse, diverse educational backgrounds, diverse experience sets. Really diversity means so much more than you and when you come into data science we have to accept the fact that there is no one and so ignoring the troll jobs 

Speaker 2: (29:35)
and looking for teams where you see someone and you say that that that's me in five years or very, very similar to me right now. Those are the teams you want to be a part of and it's harder to find those so you'll find far fewer of them, but when you do find them, there'll be more meaningful. As far as your prospects are getting into the shield, there'll be a whole lot easier. Lower barriers 

Speaker 1: (30:06)
Check out our free open mastermind Slack channel, the artists of data science loft at art of data science, loft.slack.com I'll keep you posted on the biweekly open office hours that I'll be hosting. And it's a great environment and community for all of us to talk all things, data science look forward to seeing you there. 
Speaker 1
Kind of on that same barrier to entry type of wave. What are some challenges that a notebook, quote unquote notebook data scientists face when it comes time to productionalize a model. And do you have any tips for them to overcome those hurdles? 

Speaker 2: (30:48)
I get hate every time I say this, but learn Java, just a, I mean please learn Java or CC++, please do not just be a Python developer. Your plan's amazing. Ours. Amazing. It makes life so easy and all of these tools, right? Truly amazing. And I think with the community has done is incredible. Have less five years. You look at tools that have been created and just given away for free. There is no other place for this happening in the same way. So for cybersecurity, cyber security, data science, the sharing and spreading information, but they're wonderful. However, you are going to be creating custom if you are doing import from on a daily basis. That has to be a red flag because a business which creates a model which is as simple as import from on generic data set data set, which is distinct in some way, shape or form from data. 

Speaker 2: (31:50)
Anyone else can you get, are you truly adding business stuff? And so I would say you have to look at what it takes to production, get the path to production. What does it take to productize a model? And if you don't understand the core technologies, you know sometimes it's C sharp, sometimes it's C ++ I mean sometimes you're getting down to the hardware level. That's not always necessary. Sometimes it's shovel under the covers. There's always a converge data scientists don't always need to be part of that conversion. However, they need to understand that. And in cases where they must create a highly customizable model, they need to be able to get into the weeds with the people who will be implementing it because if the two teams, again, two teams don't talk to each other using common language and a lot of time that common language is sort of the development language or the environment or so on. As you can't speak in common terms, it's hard to get something into production in the way that you intend. A lot of times it loses most of its value in translation and instead of the engineering team and the supporting teams becoming enablers, they become barriers because they're given this black box, they don't understand what it is and they just make it work in the way that they think it should work. 

Speaker 1: (33:11)
Wow. That's really valuable advice. So if you've already mastered, let's say Python or have comfortable proficiency in Python and you're looking for that new challenge when it comes to programming, then definitely go for Java. 

Speaker 2: (33:21)
Yup. I like Java. C sharp. I'm not going to offend the C sharp folks. I started doing myself. I love C sharp. Hate me windows guys. 

Speaker 1: (33:30)
Yeah. I'm uh, I'm at, I'm at a window shop myself and I do everything but they'll, they see sharp, uh, at my company. So I think I'll definitely put in the effort now to tell her that after hearing this from you. Um, yeah, I think program is just kind of a mindset. Once you know how to program in one language, it's not too much of a stretch to learn another one. Cause it's really just a mindset and a way of thinking through something. If you're logical, you should be able to, to take what you've learned in Python and draw parallels to another language pretty easily 

Speaker 2: (34:03)
if you're in engineering team. I mean I've seen machine learning models in Python where there were two comments in several thousand lines of code. If you do that in the traditional engineering team, you know, an ax murderer will come to your house. It's just Andrew career's developer. You learn structure in a way that you don't as a data scientist. And so it's worth doing and spend some time with the development. Oh, 100% worth it. 

Speaker 1: (34:31)
Good comment on comments because when I'm reviewing, uh, profiles for, you know, whatever job set up, uh, I'm interviewing for, first thing I look for when I review a candidates profile is if they have documented and commented their code. If not, that's an indication you had. This guy's going to be a fricking nightmare to work with. 

Speaker 2: (34:52)
So true. That's why I said the whole ax murderer things. So there's the meme that goes around, you know, comments or code like the person's and replace you or maintain it is an ax murderer and they know where you live. 

Speaker 1: (35:04)
Uh, so speaking of like technologies and tech stacks, what cloud technology should people pick up prior to breaking into the field? Or is this something they should even focus on if they're just looking to land their first role 

Speaker 2: (35:18)
Again with all love and respect from Microsoft world, which is where I came from and will never forget what they did for me. Amazon, Amazon, Amazon. I was like, you know, Google's infrastructure is amazing too. Microsoft's infrastructure and cloud infrastructure is amazing. I love both of them, but Amazon just so ubiquitous, so ubiquitous that it's [inaudible]. It's almost a must have. It really has become sort of everyone's go to Docker is another one to sort of slap on together. Amazon and Docker together. Just being able to build up environments, you know, from a dev ops perspective, if you have that minimal understanding, Amazon, the Amazon cloud environment and using Docker to spin up environments and quickly build environments, cloud, you know those two, you're pretty much good. 

Speaker 1: (36:04)
So before we jump into our lightning round here, I'm want to ask this final question. What's the one thing you want people to learn from your story? 

Speaker 2: (36:12)
You can legitimately come from anywhere. I mean, I was during college, I was installing servers, crimping cables, plugging stuff in, plucking things, you know, I was doing the simplest things possible. I spend time in retail, you know, to get myself through school. You can get in this field from anywhere. 

Speaker 1: (36:37)
So let's go ahead and jump into our lightning round question one. You can, you can do a contrarian point of view here as well. It's okay. Python or R? 

Speaker 2: (36:46)
Java. 

Speaker 1: (36:47)
There you go. That. Yeah, there's always a third door. Well, what's your data science? Super power. 

Speaker 2: (39:57)
Whoo, I forgot. That's a good question. It assigned super power as being obvious. It sounds like a stupid superpower, but every once in a while just saying something obvious or doing something obvious. A lot of times I use the simplest possible model and it gives me the simplest possible result. And I say, well, you know, here's what the model says. Why didn't we think of that? And those are powerful insights I provided value because we're not stuffing our Tony captain obvious. I was a superhero in data science. 

Speaker 1: (37:29)
Yeah, there's definitely elegance in parsimony. Um, what would you say is, I mean this is without context, but what would you say is your favorite algorithm for regression and your favorite algorithm for classification? 

Speaker 2: (37:40)
Oh God, I don't have one. I think if you start saying I have a favorite one or a go to one, things fall apart there. No, no favorites. Never, never, never, never. No favorites. 

Speaker 1: (37:51)
There you go. I see. So what's the number one book you would recommend our audience read and your most impactful takeaway from it? 

Speaker 2: (37:59)
Uh, thinking in bets was what got me started. It's, it is a book about poker, but it's not about poker. It's the handbook for data science and it's disguised so nicely and it's so well told 

Speaker 1: (38:13)
God, that book is fucking amazing. That book changed my life. That bayesian psychology, like just, Oh my God, that book really changed my life, dude. Cause I mean I, I read it at a point where I was trying to transition jobs and you know, I had interviews coming in and I would always just get pumped up and hyped up for a particular outcome for an interview. But then once I started thinking about it, I was like, okay, well you know, based on the situation, based on the context, this interview probably has a 5% probability of resulting in a job offer. This interview probably has like a 15 20% chance of resulting in a job offer. It just made  quelled the anxiety I think a little bit. But yeah, that book is super powerful man. Great. Great book. Yeah. So I bet you're out there crushing it at the poker tables and Reno. 

Speaker 2: (39:04)
My only bit of advice for casino is if you see somebody who does machine learning dilemma, the door lock it.

Speaker 1: (39:05)
Yeah. So certifications or self study? 

Speaker 2: (39:19)
Uh, both. I don't know. Not both. What works. I mean what you kind of said it, thinking in bets, you know, and it's kind of an offshoot of thinking in bets. Don't play to your weaknesses. You know, when you look at a particular course of action, self study or certificate, think to yourself, what is the probability that this is going to lead me? Choose something better. And part of that equation is how well do you do it? And how much you get out of it. And if you don't do it well, it doesn't play to your strengths. Don't do it. Don't worry about the stuff that you don't concentrate on the stuff that you bet you do well and that you don't have to put a lot of effort. And so that really will answer your certification. Some certifications you're gonna look at and go, Oh yeah, definitely do it. We're going to look at some self study materials and go do it. 
Speaker 1
What motivates you? 
Speaker 2
Lots of things motivate me. You know, it's interesting. It's terrible. Aye feel horrible for everyone who's impacted by it. But I think it's impacting all of us. And what's motivating me is, it's been a long two weeks, but motivated me two weeks ago I would say was seeing something of value, get to production, seeing 

Speaker 2: (40:38)
something tangible, something you built, something that was, you know, sort of recognizes value cause it's an external motivation rather than an internal motivation. But now two weeks later the world's change. What motivates me now is, and I'm coming to terms with it, I don't want to do a lot of the things that I used to do. So the grinding kind of work, I want to do things that help that are more meaningful to more people and I'm less concerned about, you know, helping the 95% more concerned about helping the people who I can bring the most impact to and health in their companies that I can bring the most. I think that's what's changed over the last for me is looking more at outcomes in a way that cards are different. And so being more internally motivated, the potential that I could make a huge difference for a company or that I could teach a group of people something important. Right It's a big change. And like I said, it's been a long two weeks. Yeah. But it's been a long two weeks for all of us.
Speaker 1: (41:44)
That's awesome. Has very, very eloquently put there. Um, and, and we're kind of off topic here a little bit, but what do you think is going to be the biggest societal impact from, from all the stuff that we're seeing with COVID

Speaker 2: (42:04)
Another thing I've been thinking about, so two weeks, you know, and I, my phone started ringing at the beginning of February and I knew something was going on. I didn't understand it was cold and that was really, uh, you know, I guess blinded by what I was doing and pigeonholing into data, that sort of thing. I was blinded by what I was working on. But like I said, I started getting the calls back and they've changed over the last and I think we are fundamentally altering sort of the landscape of how we interact with technology, how much we trust technology, how we interact with each other and how we trust each other. And there are some really dark roads, but I think the one hope we have right now is understanding each other in a better way. Bringing each other together by listening to each other's stories, personalizing what everyone is going through, those one-on-one stories or the things that we can connect about and realizing that what has happened has left people. A lot of people without hope and without those two are so closely linked that our future has a lot to do with who and what we turn to going forward for hope and purpose. That's going to be the biggest change we see in six months and two years from now is we are now making completely different decisions about what we look to for hope and for purpose. And that's going to change the way that we interact with technology and the way that we trust technology. 

Speaker 1: (43:45)
Dave, very insightful. Thank you for that. Speaking of connection, how can people connect with you? 

Speaker 2: (43:52)
You can catch me on Twitter, the underscore of shisha, you can catch me on LinkedIn. Um, okay. That's about the two easiest ways to connect with me. LinkedIn profiles, my email. You can email if you want to. I try to make myself accessible. Mmm. And it's interesting, very few people reach out, uh, aside from say, Hey, what's up? Or can we connect? I want to follow you, but I'm completely open to people that want to reach out and understand a little bit about where the journey's going and companies that are confused right now. Like  well, I guess it started in February. I've been getting a lot of outreach from companies, particularly bigger companies, but I want the smaller businesses to know. I understand how bad is this for you and startups, especially startups and small businesses reach out to me now. I'm happy to help. I'm not gonna, you know, an hour phone call. Just call me because we need to be there for each other right now. And if you have questions about machine learning or if you just have questions about how you know technology is going to change or how you could use technology to reach out to me. 

Speaker 1: (45:03)
Awesome. Andy, I'll leave a link to your Twitter and your LinkedIn in the show notes. And for anyone who's listening, I will leave a link to the book thinking in bets by Annie Duke. Excellent. Really, I should check it out. And Ben, thank you so, so much for sharing your thoughts with me and taking time out of your schedule. I know there's so much here that a lot of people are going to learn a lot from him, so I thank you very much for taking time out of your schedule and being here with me.
Speaker 2
 Thank you for having me I really liked it.
Speaker 1
Hoping to bring you back later in future.
Speaker 3
[inaudible].