SDS 245: Knowing What You Need to Know With Data Science

Podcast Guest: Luis Blanco

March 20, 2019

I had an incredible chat with Luis Blanco about the importance of data throughout a career and for companies and positions that don’t automatically have anything to do with data…at first.

About Luis Blanco
Luis Blanco is the VP of Product Compliance and Strategy at Ellie Mae but has spent 20 years of work in data science without ever having “data” in his title.
Overview
Luis has had a vast and winding path in the data science world. He started his career while a full time student in college in Mexico City. His work at American Express began as an analyst and took him 10 years later to the Director of Risk Operations after he learned the value of data science. Luis was positioned perfectly in American Express to be on the edge of data science, discovering its use as it was a budding industry. During his time, Luis worked across disciplines and across skill sets, managing teams that had to directly interact with teams that may not normally work together so fluidly, such as marketing.
Ultimately, Luis says a data scientist’s job and pathway towards progressing in their field is simply helping people solve problems.
After 14 years, Luis, who loved his time at AMEX, felt that his thought process was constrained. He began looking at the practices and ideas of other card companies, like Discover. Despite it being a large organization, Luis welcomed every single hire class and lead townhalls in the company. He organized a team of data scientists and analysts for the company who were tasked, first, with trying to find out: where is the data? It existed, but the first four months of the team’s work involved phone calls and emails to track down where the data was stored. By the second year, they “started doing cool stuff.” But it was a good lesson for Luis in his team that data science isn’t always writing code but can also be about making connections, reaching out, and engaging in the human element of data.
Looking at leadership vs. practitioner, Luis considers the difference this: practitioners are honing their hard skills while leaders are being aware of the data science world at large. What are the new applications of AI? What’s the progression of machine learning? Data science leaders need to be on top of this. While Luis’s title is not technically related to data science, he looks at the data science world at large to see how it’s affecting and moving other industries forward. Luis believes any executive needs to be aware of data science, machine learning, and AI to some extent or they “won’t be very good at their job.” Not 100% of all the decisions in the corporate world are made with facts behind them. It’s shocking, but it’s true. It’s as simple as beginning to have fact based decision making and going from there. One of the worst places to be is not knowing what you don’t know.
No matter what space you work in, data is one of the most important assets. Data scientists should make data as part of their value proposition, something the company can’t exist or progress without. When you realize what you don’t know, as a company or as a professional, you shouldn’t ignore it. Instead, dive into it.
In this episode you will learn:
  • DataScienceGO [4:40]
  • Luis’s 20 journey in data science [7:25]
  • Leader vs. practitioner in data science [34:00]
  • Do executives need to be aware of data science? [41:04]
  • B2C vs B2B spaces [45:42]
  • Going back to school [50:00]
Items mentioned in this podcast:
Follow Luis
Episode Transcript

Podcast Transcript

Kirill Eremenko: This is episode number 245 with Seasoned Executive Luis Blanco.

Kirill Eremenko: Welcome to the SuperDataScience Podcast. My name is Kirill Eremenko, Data Science Coach and Lifestyle Entrepreneur. And each week we bring you inspiring people and ideas to help you build your successful career in data science. Thanks for being here today and now let’s make the complex simple.
Kirill Eremenko: Welcome back to the SuperDataScience Podcast ladies and gentlemen. Super excited to have you back here on the show, because today we have a seasoned executive, a Vice President of Product Compliance and Strategy at Ellie Mae, Luis Blanco joining us for the show. I got off the phone with Luis just a few hours earlier and it was an amazing podcast, it was a blast.
Kirill Eremenko: What you will hear in this podcast is Luis will actually walk us through his whole career on how it unraveled. And wait for it, every single time he went to a new step in his career, you will learn a valuable lesson, a valuable takeaway. This is a career of over 20 years packed into a one hour podcast where you can get insights at whatever level you are. Whether you’re an analyst looking to build a career and eventually become an executive one day. Or whether you’re already an executive and looking to better manage your analysts and fill in those gaps in data science knowledge and leadership that you might have right now. In addition to all of that we talk about very interesting topics such as planning your skills rather than jobs for your career development. Fact based decision making cultures and how to nurture, how to create and nurture them, cross departmental work and sharing models between different departments such as risk management and marketing and much, much more. Lots and lots of valuable insights you’ll get from this podcast. I can’t wait to get a start at it, so without further ado, I bring to you Seasoned Executive Luis Blanco.
Kirill Eremenko: Welcome back to the SuperDataScience Podcast ladies and gentlemen. Today I’ve got a super exciting guest joining us from San Diego, Luis Blanco. Luis, welcome to the show. How are you going today?
Luis Blanco: Thank you Kirill, very excited to be here and being able to chat about all the cool things data science. Hopefully being able to give something useful to all of our community out there.
Kirill Eremenko: Fantastic, thank you so much for joining us. You mentioned it’s pretty cold in San Diego today, what’s happening there?
Luis Blanco: You know what I think that it’s probably climate change. No, we have a very, very wet winter here and that is making temperatures drop. Yeah, I mean we didn’t even make it to 60 degrees Fahrenheit today which I think it’s like 14 degrees celsius for the rest of the world. Yeah, I mean I have a friend that is visiting in, he’s a little bit disappointed that his San Diego vacation included a full jacket and shoes.
Kirill Eremenko: Yeah, it was really funny like when we were chatting just now before the podcast when we had video. When I saw you in those two or three jackets that you’re wearing, I thought you were in Alaska or something like that.
Luis Blanco: Well that is that and the fact that I’ve been sick for the last couple of weeks though. I apologize in advance for all of the listeners that I may be coughing a little bit here and there. I’ll try to minimize it, but what do you do?
Kirill Eremenko: No worries. Thank you so much for taking the time and array through it. Luis, we met at DataScienceGO particularly at DataScienceGOX. For those of our listeners who don’t know this is a current event. We have DataScienceGO which is a conference for data scientists. This year it’s starting in September 27-29 in San Diego. This GO X is a co-current one day event specifically for executives where they can get together and have a forum, talking about data science leadership and strategy. Tell us just quickly how did you find DSGO X and what kind of value did you get out of it?
Luis Blanco: Yeah, absolutely. You know I was honored to be part of the DSGO X team and kudos to all of the folks that put together the conference. It was one of the best that I’ve attended in a very, very long time. The smaller setting was very, very useful. What I can tell what I got the most out of it was relationships that you can build in that smaller setting. The conversations that we had were deeper than you could in almost any other forum. You guys did a really good job on keeping really flexible in the way that we manage time. Some of these conversations you may remember we went on for a couple of hours. Both the relationships that I was able to build as part of that as well as those in depth, they were rich conversations certainly set DSGO X apart.
Kirill Eremenko: That’s awesome. You just think that you mentioned flexible right away I thought of how during the executive dinner we were all switching seats. How did you find that exercise?
Luis Blanco: I think it was very good proposal. This way you have the chance to talk to other folks. I think that that is a principle that you guys should completely leverage in the future, mixing things off as I don’t want to call it forced, but in a way to push a lot of folks out of our comfort zone. Particularly in our community let’s face it, we have a large percentage of folks that tend to be more introvert than extrovert. Unless put in this type of context maybe we tend to not be as open. I think that worked out very, very well. I would say go a little further you know, spice things up a little bit more.
Kirill Eremenko: Thank you. Amazing feedback and for all you executives and directors, business owners listening out there, check out DSGO X happening this year in September. Luis, like I don’t even know where to start with your career. You have such a diverse background. You spent over 14 years at Amex, now you’ve moved on to very interesting senior executive roles. You are also involved in data science, so many interesting things to talk about. Maybe to get us started, can you give us a quick overview, a brief background of who is Luis Blanco and where has your career taken you?
Luis Blanco: Yup, absolutely, thanks Kirill. Like you say careers, when I was in the earlier part of mine I thought of it as a linear function, taking you from point A to point Z via B, C, D, E and F. Then the further that it came along I realized that there’s nothing further from the truth. I still plan though methodically, but instead of thinking of next step in terms of jobs, I think more in terms of skills that I need to continue to develop. Then try to match opportunities to that skill in that path. I know that I haven’t planned for it, I may have missed critical opportunities just because I wasn’t thinking of them.
Luis Blanco: I started my career with American Express many, many years ago while I was attending college in Mexico city. My undergrad is in business management and finance. I wanted to get a head start in my professional career, so I decided to get a full time job while I was attending full time college. That was interesting. It was very busy, but it paid off. I started working in the credit operations function. Within a couple years I get promoted a couple of times and before I graduated I was already leading a team of seven people locally and then I had direct reports throughout the rest of Latin America.
Kirill Eremenko: Wow! Before you even graduated?
Luis Blanco: Before I even graduated, yeah.
Kirill Eremenko: That’s crazy.
Luis Blanco: The company was expanding, they were growing and the operation of Mexico was also growing as both as Mexico market economy grew, but also as a regional hub. I was able to make the most of that opportunity. I graduated in June of 2001 and shortly after I graduated I had the opportunity of being considered for a manager job in Amex’s international risk management group in Phoenix. That used to be, we called it the International Risk Management headquarters because that’s where we had all of the talent that helped manage credit and fraud risk for American Express. After a very thorough and very competitive process, I was lucky to get selected and that’s when I moved to the United States and started working with that International Risk Management team. It was fascinating.
Luis Blanco: This was my first true global job. I led projects as well as what we call credit policy for late part of the credit card lifecycle throughout Latin America. Then I had a chance to also interact with Japan, Asia, Pacific and Australia. It is through those gigs that I had a chance to visit Sydney that we were talking about earlier. That was my first purview, very, very small purview, but about the value of data science.
Kirill Eremenko: Interesting.
Luis Blanco: I know that that is 17 years ago, but credit risk is one of the areas that used advanced statistics. What we now call the entire data science lifecycle in making those credit decisions, the basic credit decision could you lend and how much do you lend them for. Me coming from an operational function, for me it was all about project management efficiency so on and so forth. This was the first time that I saw just an incredible amount of value that was generated by doing things differently. One of my first bigger projects was to implement business rules that will determine what type of accounts do we send to a particular type of vendor. This is one step before predictive models, but it had already changed dramatically with the World Trade International. Before I left that job we were implementing our first predictive model.
Kirill Eremenko: You were doing data science as it started becoming a thing? As it was becoming born, you were already dabbling in it, so you’ve been there from the very start?
Luis Blanco: Yeah, exactly. I mean for the international markets that was that and it was, till I moved to American Express’s headquarters in New York city a few years later that now I saw what was the leading edge of data science in the mid thousands. I moved to New York in 2004. 
Kirill Eremenko: Luis may I just inrerrupt you? If anybody is at their computer right now, open up Luis’s LinkedIn, it’s so fun to follow along, because you have this all written out in your LinkedIn, all these roles very carefully articulated. I’m just watching your LinkedIn as you speak and following along. It’s like I’m part of a movie with a timeline, pretty cool. Sorry to interrupt.
Luis Blanco: Thank you. No, no worries. Yeah, please go ahead. In the US risk management team, now I saw what was the state of the art in data science while we were implementing our first predictive model in international. Over there they were in the third generation of predictive models. I had the chance to work with one of the most brilliant statisticians, econometricians what now we would call data scientists in the world.
Luis Blanco: I mean they were breaking ground with, I remember one of the models that I implemented was the first use of the equivalent of decision trees today. This is on all their technology, but part of our challenge was just to be able to run them through our dated servers, non cloud enabled any of those things existed back then. At that point one of our projects it was just incredible. By just enhancing the model, improving the predictive power we were able to put about five million dollars to the PNL of that year. Which was a relatively small effort, but it yielded incredible efficiencies down the road. I was doing that happily living in New York and my previous leader reached out to me and they were growing the risk management functions in the key markets for American Express International. He said that if Mexico is a key strategic market for Amex and he asked me if I was okay to move back home on a temporary assignment. We did just that and my wife that was a year that we got married, so it was a lot of stuff happening at the same time.
Luis Blanco: I didn’t know that the financial crisis was coming and it hit the Mexican economy particularly hard because of the economic dependency in the US as well as the financial system was back in the day I think that 90% of all the assets were owned by banks with foreign headquarters. You know HSBC, City Bank, Royal Bank of Canada all of them are big, big investments in Mexican banking system. When the liquidity crush happened, they took the liquidity out of the Mexican market to put it back to their home countries. That just exacerbated the crisis.
Luis Blanco: From a portfolio management perspective and particularly when it comes down to data science, we needed to come up with all the ways that we could turn around the profitability of a one billion dollar credit card business and personal loans. We also needed to integrate new players like for example the main credit bureau in Mexico was going through a big transformation. That meant that we needed to update our data intake mechanism and then fully integrate it with all of our different systems so that we can continue to run our business, all of our predictive models and so on and so forth. That was a very, very exciting time. It was stressful time because we just saw our, it’s called the loan provision or that amount that banks have to set aside for a loss in future. We saw that going off as the rest of the industry, but we needed to make sure that we had a good plan. I did that for about …
Kirill Eremenko: Interesting there is that even though you were the director of risk management, your title doesn’t say data in there. Before 20 years ago, director of risk management wouldn’t be thinking about how do we intake data, how do we make sure that we are aligned in terms of the fields that we have, the fields that our suppliers provided, our models are working well. Even though you’re the director of risk management, your role involves a lot to do with data, is that about right?
Luis Blanco: That is absolutely right. Kirill, a lot of what my team and I did was precisely to A: ensure that what we now would call our data infrastructure and the architecture was sound. That we needed to develop some new pieces. Of course, we worked with our technology partners. We didn’t own the servers and the code, but we did own the strategy, so yeah, absolutely. I mean a lot of it was during that work. Then after the fact, once you have built all the pipeline, now it is on the decisions being made on that data. Our decisions were who do we lend? How much do we lend? After that fact, that relationship continues, how do we continue managing that credit risk?
Luis Blanco: Another part that was fascinating about that job is that I worked very closely with marketing in managing the portfolio. One of the things that we did back then was, marketing had their own predictive models, and they were looking at revenue growth as the objective function. We helped them to enhance their models to now look at the whole profitability cycle, not just revenue, but also look at profitability. Any data, a big, big difference.
Luis Blanco: Financial crisis are tough because it seems that the world is coming to an end. The truth of the matter is that it doesn’t across our hundreds of thousands. As a matter of fact we had millions of customers in Mexico. There were segments that were risky, and we needed to manage them prudently. But there were segments of the portfolio that it was worth continue investing, actually investing more so that we can gain more of their share of wallet and their loyalty. That was a very cool part about the job.
Kirill Eremenko: Sorry if I may add here, it might seem as a not something that stands out in this podcast, but I think there’s a very important point that you mentioned. That in your risk department where you are the director of risk management, you were creating models to make sure that you give the right people the right loans, the right interest rates. You managed the risk of the company. In the marketing department on the other hand, they have their own models where they segment their customers to send them the right offers. They go out to maybe get geo demographic data and reach out to new customers and so on. The KPIs might be different, their KPI might be, okay, let’s get as many customers as possible, potential leads or let’s get as much revenue as possible, predicted revenue.
Kirill Eremenko: Let’s say they bring those customers in, they market them, the American Express in this case or whatever other company, they increase the brand awareness and the customers come to. Then you still apply your models to decide who you will service and who you will not. It’s a great example of where in large organizations’ departments should be working together, there should be some cross departmental work around data science so that the models that they’re using and the models that you’re using are not contradicting each other. They’re actually working in unison together to bring benefit as you said in the form of profitability to the company overall.
Luis Blanco: Yup, I completely agree with you and one of the learnings was precisely that. After I left Amex, there were a lot of efforts being done in trying to identify this type of analytical work groups across the company. Put them maybe not formally from our reporting perspective, but create a lot of relationships so that people can share their knowledge. There was certainly an effort to share infrastructure across multiple analytical teams.
Kirill Eremenko: Yeah, that’s fantastic, I think it’s a very important point. Thank you for bringing that up.
Luis Blanco: Yeah, and then as I was doing all of that, the crisis had subsided. I’ve mentioned how from a career planning perspective, you know I’m not necessarily looking for the next job, but skills. As a director at Amex I led a small group of highly, highly trained professionals and everything else we did throughout the matrix. I’ve realized that leading a large organization when you have, you know you don’t have the luxury of seeing your entire team every single day. You cannot leverage just that type of friendly relationship to make things happen. That was something that I would need to learn how to do. I was very lucky that there was an opening to lead the risk operations team in continental Europe based out of Madrid in Spain. This was moving after eight years risk management really tailoring a lot of my analytical skills. This was more to continue polishing my leadership skills.
Luis Blanco: My wife and I and our Cocker Spaniel decided to relocate to Spain on a temporary basis, then lead that organization. It’s about 250 people. I had at some point seven direct reports ended up being five managers and then supervisors that my consumer facing people I reported to across multiple markets in Europe. That was a fascinating experience of having to understand the cultural differences of our consumers. For them how they view the American Express brand was very different. Even though we stand for the same set of corporate values, the perception of those were different. Also, it was very diverse group of individuals that I had responsibility for. It was a wonderful leadership experience.
Luis Blanco: I had also the chance to go back to school and get my MBA at the IE Business Schools one of the top business schools in the world which was, that on its own an incredible experience. I was able to meet some fantastic folks.
Kirill Eremenko: Wonderful.
Luis Blanco: I apologize. I did that for two and a half years. This was a lot of blocking and tackling, not a lot of data science per se, but the difference … I mean it’s funny how I ended up being there and I didn’t have enough data to make the decisions that I needed to lead a large organization. Our reporting capabilities were basic and it’s funny how I couldn’t get a straight answer data based, that based on what is it what we do. We have 250 people, our scorecard is very high level. What is it what we do? I was able to hire a data scientist who was an actuary by trade, and help me start identifying the places that we could get the data so that we can answer that very, very simple question.
Kirill Eremenko: Fantastic. You move into this role managing 250 people. You see that you don’t have the data and your first move is hire a data scientist, let’s sort this out. There’s a great saying, you cannot manage what you don’t measure.
Luis Blanco: Exactly.
Kirill Eremenko: You need to be measuring, that’s a really cool approach.
Luis Blanco: Yup. I totally agree with you Kirill. It was not only because of my passion for data science. I mean that is true, but it was to be able to do my job more effectively. For our aspiring data scientists or our data scientists that they are at this point in their career, that they are like, “How do I get to the next step? How do I get more responsibility?” The most important thing is, help people solve a problem.
Luis Blanco: Sometimes it’s easy to get lost in the technicalities of data regardless of where you are. If you’re building the pipelines or if you’re on the analysis side of data science, it’s sometimes easy to don’t see the forest by looking at the trees. In this case my data scientist was critical in helping me answer that question, well what the hell do we do across all of these different functions? Most importantly it gave me the foundation of our strategy for the next couple of years. That was key.
Luis Blanco: Fast forward a few years and for the first time in my entire career, as I graduate for MBA I realized that while I’m very grateful for the 14 years at Amex and all of my leadership team and my team members. I’ve also realized that my thinking was constrained. I’ve only worked for one company, I’ve only known one way of doing things. For the first time I start talking to other companies, other recruiters and Discover has this very interesting role based out of the operation center in Salt Lake. To lead an organization of 500 people in their customer service organization.
Luis Blanco: The cool thing was that it also had revenue generation responsibilities. That really caught my eye. I still had a chance to lead a large organization, so keep on developing that skill. Now I’m also to start contributing to the top line of the PNL.
Kirill Eremenko: Just quickly, just cover for those who don’t know, Discover is another type of credit card, right?
Luis Blanco: Oh great point, yes. Discover, we operate under the Discover brand, issuing credit cards in the US. Internationally we also payment network and we operate the Diners Club Network.
Kirill Eremenko: Oh Diners Club, yeah. I had one of those cards that was cool.
Luis Blanco: Exactly. For our listeners outside of the US, you might be familiar with the Diners Club Network. The Discover folks they did a really cool thing. Payment networks are not easy things to build. Growing internationally, it’s not an easy thing to do. They identified that they could become the network of networks, so the growth for the Discover Network International has been actually the connect networks across them. If you’re a local payment network in China and you have a percentage of your credit card holders that travel internationally, now because of the deal, now the cards can be used anywhere that Discover and Diners logo is. It’s a very interesting strategy.
Kirill Eremenko: Thank you.
Luis Blanco: Yeah, I mean in this case a large organization, very customer oriented. Discover is a fantastic company, great leadership, but came back to a similar problem that I had in my previous job. I lead a large organization, it’s 500 people, there’s no way I can talk to all of them every day.
Kirill Eremenko: You probably didn’t even know all of them.
Luis Blanco: I was very lucky that our company was very focused on the people. I welcomed every single hiring class that we had, and I had the luxury of having town halls with my team. It took me a little while to meet all of them when I first joined, but after six months I can tell you that I met everybody that worked in my organization.
Kirill Eremenko: Nice, that’s awesome.
Luis Blanco: Experience life base, you need to rely on data. Our reporting capabilities were an Excel sheet with a couple of groupings that you can open up and that was it. After couple of years of bugging my leader and him asking questions and me not being able to answer them, we agreed to invest in analytics function. I built the what we call the field analytics group, which was actually a team of data scientist and analyst to being able to answer questions related to our operation. This was a very interesting journey, I’m sure that a lot of our data scientists go through a very similar journey. 
Luis Blanco: The first question is, where is the data? I know that we generate rims of data every single day across multiple of our operational systems, but where is that data stored? I hired my first employee and he thought that he was going to be in Python, writing cool code from day one and the first four months it was a lot of phone calls, a lot of emails and actually trying to identify that. Then the second part was identifying what was our analytical environment. Large companies sometimes they have their own and in our case Discover we were married with SaaS which provided a lot of great opportunities, but it also had some limitations. We needed to figure out what we could do within the SaaS environment.
Luis Blanco: Then it was not until the second year that we actually started doing cool stuff. For colleagues out there, that journey is very natural. It’s part of the cost of innovation. Is that sometimes it’s not all going to be 12 hours sprints in building code. Sometimes it is reaching out to people and getting their help to understand where things work. It was in one of these instances that we’re trying to solve a business problem that I need to beg and borrow time from our central of excellence of analytics that I realized that I needed to learn more stats and data science. I decided to go back to school and enrolled at originally it was the master in predictive analytics at Northwestern University. Since then, it has evolved into a master’s in data science.
Luis Blanco: I’m 65% in completing that degree. I had to put it on hold a little bit easier with well last year with changes in jobs and what have you. That’s where I got a lot of the technical skills. Since then moved from Discover went to CoreLogic to lead the business risk team for credit services. Just four months ago I joined Ellie Mae to lead their product plans and strategy function.
Kirill Eremenko: Fantastic role. Congratulations on the recent move in your career and thank you so much for this overview it was really cool to follow along. I think in every time you talked about a new role that you had in your career, every time I got a takeaway. I think our listeners as well. We’ve got an interesting takeaway from your experience, so that’s a very cool overview. Thank you so much.
Luis Blanco: Absolutely.
Kirill Eremenko: Okay, so you’ve gradually progressed from doing analytics work and starting to lead a team to then being an executive and running organizations and running teams of 250, 500 people. How does it feel to progress in your career like that? When you look back on how you were starting out, what are the main differences between being a leader of an organization and specifically I’m talking about like in terms of data science. How do you view data science as a leader versus how you used to view it when you were a practitioner mostly in this?
Luis Blanco: Right. You know as a practitioner it is really important to hone your technical skills. That was probably depending on new job and what have you, but it would be 20 to 30% of the time. It was easy to sometimes forget that you need to keep on doing it. From a leadership perspective it is important to, now that time shifts, instead of honing in your own hard skills, it’s also to know what is happening out in the street. That way it was so important for me to go to a conference DataScienceGO to find out what are people talking about, what is the latest and greatest. The truth of the matter is, it’s not only about what is next analytical software.
Luis Blanco: When R now gets replaced by Python and Python is going to be replaced by something else, we know it already, we just don’t know what it is. Most importantly what are the new applications of AI? Or how is machine learning changing some different businesses? It’s funny, my title doesn’t say anything about data science right?
Kirill Eremenko: Yeah.
Luis Blanco: I’m not a chief data science officer, but a lot of my job is precisely to find ways of making our products more efficient in managing our own operations in a more effective way. Data science helps in bridging those gaps. In summary the way they would think about Kirill is going from needing to learn and be an absolute expert in the technical to now have to know a broader sense. Even if you don’t go as deep, I certainly need to know broader. Not only how in my previous life how do you apply machine learning algorithms in the credit space? Now I need to understand how AI is changing entire industries across the board to see if there’s anything out there that we can then apply to our own business problems.
Kirill Eremenko: Gotcha. And do you have any examples like recent examples of something that in this broad approach to data science and what’s going on in this world, any examples where you learned something that has been or might be helpful for your organization? Of course, I don’t want to ask you to disclose any trade secrets or anything like that. Maybe something that was helpful for yourself.
Luis Blanco: Yeah, I mean I could tell you in Ellie Mae one of our value propositions is related to compliance, the real estate and in particular the mortgage space is highly regulated to the United States. There is a lot of regulation that is being looked upon. What we are looking at is how is the so called wreck tech evolving in other spaces that are not related to mortgage and how we can apply it. That is certainly the early stages. I can tell you back in my role at Discover, what we were doing was applying machine learning algorithms as part of predictive modeling to being able to predict multiple steps in the consumer life cycle.
Luis Blanco: The last project that I did we were marketing team, we had balance transfers, it’s a pretty big revenue growth in the credit card industry here in the United States. We were working with them leveraging my teams technical expertise and their knowledge of marketing and how we can enhance our campaign models leveraging state of the art machine learning algorithms.
Kirill Eremenko: Well okay. Being up to date with what’s going on in the world helps enhance your knowledge and helps you guide you in the right direction in exploring new things. Whether they’re right or wrong, but sometimes of course you fail when you’re exploring new things. Eventually you find a right solution, things that can enhance the work that you do.
Luis Blanco: Yup, absolutely. I mean in DataScienceGO for example it was in one of the sessions with Gabriela.
Kirill Eremenko: Oh yeah.
Luis Blanco: From IBM. As she was talking we had a problem. You know I cannot go into a lot of detail because of the nature of it at CoreLogic, but it just hit me that we could do a similar approach to what she was talking about to solve some of my problems. After the conference, I get back to the office, talked to a few of our folks so that they can start exploring those type of solutions. It was not exactly what she was talking about, but sometimes that’s the way ideas flow. Something connected the dots between what she was very kindly sharing with the rest of us and the problems that we had at work that it was like, oh okay. This is something that we should approach this other way. Yeah, innovation comes from interesting places.
Kirill Eremenko: That’s so cool. Was that [inaudible 00:40:16] it was her talk about like deep learning models on a napkin or something like that, like very quick deep learning models?
Luis Blanco: Yup. She was particularly talking about the ease of deployment.
Kirill Eremenko: Oh yeah.
Luis Blanco: When that struck when she shows her GitHub and everything.
Kirill Eremenko: That’s so cool. That’s awesome. Did you manage to meet Gabriela at DSGO?
Luis Blanco: I did, yeah we had a few words. I got a chance to introduce myself and we’ve been in touch since then.
Kirill Eremenko: That’s so cool. Awesome. Well thanks for sharing that’s a really cool example. Like the way you described your career like there’s so many questions that I was writing down as you were talking. You just like opened up this whole field for different interesting questions. The next one I wanted to ask you was, like something you mentioned and we saw this a couple times. That throughout your career I don’t think even once you had the word data in your title. Nevertheless, you’ve always been working data, with data very closely. Would you say that at this stage any executive needs to be somewhat aware of data, the power of data science, artificial intelligence? If that’s the case, then what kind of tips would you have for them?
Luis Blanco: I think that they should Kirill, honestly. If they don’t they’re probably not going to be very effective at their jobs. For all of us that lead teams or lead organizations or have responsibilities on product lines and so on and so forth, number one I know that these may sound like obvious, but it’s still not 100% of all the decisions in the corporate world are made with facts. I can tell you, in the 20 years of my career, I have been shocked as how many decisions are made actually without the data based on well intended individuals that have tons of experience in that domain. But as data scientists we know that sometimes the data tells you counterintuitive fact and the decisions are typically over the long run better when you involve facts.
Luis Blanco: The first thing that is out there is, if you don’t have a culture that is fact based decision making, that’s the first thing that you could do. Once you do that, then you’ll start realizing that your data science infrastructure it’s probably lacking which then goes to the second point. It is, even if like you said data scientist is not on your title, you’re very quickly going to find out that you may need to hire one that says data scientist at the title. Unless you work at Google or some of these companies that data is in their DNA, then you’re [inaudible 00:43:22]. For the rest of us of mortals out there, it’s very likely that you’re not going to have all of the data that is required to make those decisions. That you’re going to have to go out and build it. That’s when knowing about the basic tenets of data infrastructure, knowing about statistics, you don’t need to be developing new algorithms for your own. You need to understand the basics of hypothesis testing. You need to understand certain, I don’t want to call it basic statistic knowledge, but you need to be able to tell facts between noise. I think that that is the key distinction from what I’ve seen are extraordinary executives throughout my career versus the ones that were just okay guys. 
Kirill Eremenko: Yeah. The way I see what you mentioned is, when executive, when you bring him some results and facts, some findings, the first question that I expect to hear is, “Is this statistically significant?”
Luis Blanco: Exactly.
Kirill Eremenko: Rarely that happens. Obviously I go and do my checks, before I would go and check that these are valid findings that people can rely on. What if the data scientist doesn’t do that? What if the insights you’re looking at they’re saying one thing and they based on numbers, but they’re not statistically significant. You need to know those kind of things.
Luis Blanco: That’s exactly right.
Kirill Eremenko: For sure. Another thing about data what I like, I think this is like it comes out of these things that you’ve mentioned about the importance of having data science on board is that, data allows you to know better what you don’t know. Sometimes you think you know a lot of things and so on, but there are so many things that you don’t know and until you have a data science team or that culture of I love how you said it, fact based decision making culture. To really have that culture in place, you don’t even know what you don’t know. I think that’s one of the worst places to be. It’s better to know what you don’t know than be completely oblivious.
Luis Blanco: Yup, I completely agree with you Kirill.
Kirill Eremenko: Got you. Okay, awesome. That was some great insights and another question I had was, in your roles and we talked a bit about this before the podcast, you started off in the space in the world of B2C. Predominantly as I understand it in Amex from a business stand point you’re dealing with customers from business to clients, business to retail customer. Then with time you moved more into, especially in your most recent roles at CoreLogic and Ellie Mae you were in the B2B space, so business to business. How do you find the differences? What are some of the core differences involved in data science between B2C and B2B?
Luis Blanco: It’s very interesting position that we Kirill, in both B2B environments that we deal with, data is one of the most important assets. Our customers look at Ellie Mae so that we can help them complete all of the mortgage transactions, but also to manage data from one step of the mortgage process to the next step. What we do is, we create data solutions. In that perspective it was very similar to the work that we did in Amex. It’s just that we used to do it just for one client, us. Now we create solutions for an entire host of customers across multiple segments of the mortgage origination market.
Luis Blanco: The big difference though that I see is scalability is a big, big difference. I mean Amex and Discover both have several millions of customers so they’re not a small company of any structured imagination. If you compare that versus we at Ellie Mae we service about 40% of every single mortgage that happens in the US. It’s a $1.3 trillion industry.
Kirill Eremenko: Wow!
Luis Blanco: Even at it’s busiest day the strain that we would put our infrastructure on Amex wouldn’t be compared to what we have at Ellie Mae. Then the true technicalities on how we do things matters. We offer data services to our customers and the way that we build those data services, the architecture of that matters because we could put production at risk. With that will come a whole host of problems of reliability. Certainly scale matters and in B2B obviously there are very wide industries out there. I will be surprised if even as short as five years a lot of the value that is being provided in B2B is related to the data. I mean think of with IOT in industrial, that is revolutionizing industrial engineering, both design and then servicing after that fact.
Kirill Eremenko: Wow, that’s a very deep thought. A lot of the value in business to business is going to be provided with IOT and data? Yeah, I think you’re right. It’s becoming such a, so your beacon is data right? It’s about just like adding it to your value proposition, adding it to your offering and showing businesses how they can make that a resource, an asset that they can leverage and make it better. Wow, yeah, never thought of it that way. Very cool.
Kirill Eremenko: Okay, so we talked a bit about B2B versus B2C. I was very inspired by your comment that you went back to school to do that masters of data science at Northwestern University. For somebody who’s already done all of their education and even at MBA and is already in an executive position, that’s a massive decision. It requires a lot of courage and where do you find the time? How do you even consider going back to school? How did you make that decision? What kind of inspiration can you share with other executives or are other people who’ve been in the workforce for sometime and who are considering to continue learning? Whether it is through university or whether it is online, whether it’s in other ways. What kind of impulse can you share with them?
Luis Blanco: Yeah absolutely. I mean it was not an easy decision and I’m lucky that I have a fantastic wife, extremely supportive. I have to tell you that when we started talking about it, she was like, “Again? We’re going to go through this again? What’s in the MBA card already?” We did that before we had kids. It was a conscious decision and it came from that knowledge of what you don’t know. Discover was always very, very supportive of my development. The MIT they have in the summer this like crash courses. I did one in a discipline called design of experiments, DEO. Which interestingly enough it was something very different than what I thought it was, but very, very useful.
Luis Blanco: Throughout that week I came to realize that even though my undergrad and my MBA they are analytical in nature. All of the years that I spent working in risk management, but there were areas of statistics that I didn’t know. I’ve realized that I didn’t know what I didn’t know, and I started looking for ways of trying to cover for that knowledge gap. It quickly was obvious to me that if I kept on trying to do that through more like courses this type of engagements, it was going to take me forever. I didn’t even know how to structure it, that’s what schools do, they know how to build knowledge trees and so on and so forth.
Luis Blanco: I started looking for options and the schools in Utah which is where I used to live before didn’t have anything that was close to what I needed. Starting to look at distance learning options and Northwestern University has a fantastic program. It’s deeply technical, but it’s not only a technical program. After a lot of discussion with my leader at the time and with my wife, we all agreed that it was a passion of mine. Then that this would pay off in the future. That’s how we decided to go into this again.
Kirill Eremenko: Fantastic. Basically the main kind of piece of insight I get from that is that, when you feel there’s something that you don’t know, you shouldn’t avoid that, you should dive into it and discovered what it is that you don’t know and find ways to fill that gap. Close it and actually augment your knowledge because that enlightenment will take you to the next level as an executive or as a professional. Does that sound all right?
Luis Blanco: Yup, absolutely and I used a simple rule when I look at education in general. You get knowledge, you also get network and you also get the curricular value. Depending on the choice that you take, it may not be equal across these three different areas. If you decide to go on your own on [inaudible 00:54:34] for example, you might get the knowledge. You may get an access to a network, but you certainly, the value of a formal network like you have in SuperDataScience or the ones that you do through shool alumni and that type of stuff, it’s not the exact same thing.
Luis Blanco: There are instances where the title matters. If you have an undergrad degree and you haven’t completed it, well getting that undergrad degree is going to make a big difference in your curriculum. Depending on what stage of your career you are, both your professional and academic career, there might be different options available. The decisions that I’ve made on education, I put it through this lens and make sure that I’m very objective and honest what I’m trying to achieve. Then try to find what is the best opportunity to do that.
Kirill Eremenko: Fantastic. Thank you so much Luis. We’ll leave it at that. This brings us to the end of our podcast. Thank you so much for coming on this show. It has been a huge pleasure. Before I let you go, can you please share with us what are some of the best ways for our listeners to get in touch, contact you, follow your career, maybe possibly ask you some questions if they need some guidance?
Luis Blanco: Absolutely. I mean the best way’s through LinkedIn. Go in Luis Blanco working at Ellie Mae. Send me an invited, I’ll accept it and we can continue our conversation through that means.
Kirill Eremenko: Fantastic. Thanks so much Luis and of course we’ll share the URL as well in the show notes. I have one final question for you, and I’m a bit shy because I know what you’re going to say. Nevertheless, what is a book that you can recommend to our listeners to further their careers?
Luis Blanco: Absolutely. I know that you thought that I was kidding, but I’m not. I was lucky to get a copy of your book Confident Data Skills during the conference. I went through it. Like you and I talked about it. I love the approach that you took to the data science. I mean being in grad school I have hundreds of thousands of very technical books sitting here right next to me. Let’s face it for the most part they’re very dry. It has technical content that you need to understand.
Luis Blanco: You’ve achieved very difficult objective of having a document, a book that has technical insights. You finish that book and you know how to do stuff, which is very important. This is not preaching just importance of data science and how cool it is, all of those stuff which is a plenty out there. This gives you specific skills on how to do particular things, but in a way that you end up reading the book and it’s like, “Oh you know what it’s not that bad, not that hard.” It’s actually fun to do it. You did a really, really good job there.
Kirill Eremenko: Thank you, thank you so.
Luis Blanco: Yeah, I mean Confident Data Skills by Kirill Eremenko, it’s published by Kogan Page.
Kirill Eremenko: Thank you so much Luis. I’m flattered and very honored to hear that coming from you. Yeah, as we chatted before the podcast, I was very surprised and excited that you actually had read the book, so thanks again for putting up.
Luis Blanco: Absolutely and thank you and your listeners for their time. I hope that they find this useful in any way, shape or form. Thank you Kirill for inviting me.
Kirill Eremenko: Fantastic. Thank you Luis.
Kirill Eremenko: There you have it ladies and gentlemen, what a podcast. So many insights I was personally glued to this audio and to my screen as I was walking through Luis’s LinkedIn with him and I learned a ton personally. Out of all the insights which were all amazing, I have a few favorite ones, probably three favorite ones that I identify for myself. First of all what Luis said about planning your career based on the skills you need to develop rather than the jobs that you want to have. Then working, the concept of departmental work, how departments should work together and create shared models or at least models that are not contradicting each other. For instance that example of the risk management and marketing department working together. Finally, the fact based decision making culture which is such a powerful notion.
Kirill Eremenko: I hope you enjoyed this podcast as much as I did. As usual you can get all the resources mentioned in today’s episode in the show notes which are at www.www.superdatascience.com/245 that’s www.superdatascience.com/245. Make sure to connect with Luis on LinkedIn. He’s a great person to have in your network. Of course if you’re interested in learning more about DataScienceGO or DataScienceGO X for executives which I highly recommend we’re taking 40 executives on this year. If you want to be part of that community, if you want to be part of that executive experience and make those valuable connections and network with other execs interested in data science and learn from each other and grow your executive network, then head on over to datasciencego.com and check us out there. You’ll find a specific application ban for DataScienceGO X. That’s www.datasciencego.com.
Kirill Eremenko: By the way if you know any executives, if you’re not an executive yourself and you know of some executives that might be interested, maybe your boss or maybe a friend of yours, a colleague, a family member, send them this episode and let them get these insights as well all these wonderful valuable insights that Luis shared today. On that note, thank you so much for being here today and I look forward to seeing you back here next time, until then happy analyzing.
Show All

Share on

Related Podcasts