Data Science Hangout | Matthew Montero, Gen Re | Bringing a Vision to Life
videoimage: thumbnail.jpg
Transcript#
This transcript was generated automatically and may contain errors.
Matthew Montero, Gen Re. Happy Thursday, everyone. A little bit of a change of pace today. So Rachel is on vacation. How dare her? But I'm stepping in to at least try to be a host for that Hangout. Welcome to the Data Science Hangout. If this is the first time that you're joining, welcome. We're super glad that you're here.
If you're returning, which I suspect a lot of you are, welcome back. As a reminder, this is a casual conversation and a casual environment. So if you're here and you have thoughts, we want to hear from you, no matter how much experience you have or don't have. This space and this hour is really for you.
So with all that said, I want to introduce our leader or featured leader for the day, Matthew Montero. Super glad that you're here. Maybe introduce yourself and tell us what's going on.
So my name is Matthew Montero. I head Enterprise Data Services within Gen Re on the IT side, actually, so it might be a little bit different where we actually provide services for the business. So whether that may be database services or data lake services, data science services, which I think a lot of you guys would be interested in, and data reporting services. And then with that, it's all about provide systems and applications that are hard-earned by IT, but then also in a self-service way as well, meaning that here's a platform that the business can build on or improve on, and we support to make sure that they'll stay up and running.
My background actually is actuarial, about 12 years back, I graduated college, got a hold of being an actuary, doing all those exams. I landed into a job where I got to do both actuarial work, but then got a little flavor analytics. And from then, I grew to overseeing a team, building a team from ground up and overseeing that team. Eventually, I had a little time in pharmaceuticals, which I also was building a medical team to solve the clinical trial automation challenge. But then from there, I moved on to here, to eventually being a whole data organization reporting to the CTO.
From individual contributor to senior leadership
So even from the beginning, I was always interested in data, and also the process to engineer that data to actually get some insights from it. So being an actor is always very analytical and trying to solve that, so that's what my individual contributor is like, let me build some applications, or let me try to find ways to simplify that process, or at least automate the process from there.
But then when I was maybe five, six years into that space, I was like, well, I want to start leading teams, I want to start building, working with others and mentor others to start being able to do that. So that's when I started becoming more of a team lead or a technical lead.
But once I got into pharmaceuticals, it was very different because the company was a lot bigger, and then I had a pleasure to having a manager who was very focused on vision. And so that's when I really got the feel for like, once you have a vision, and you can sell that vision, then you have the ability to build as big as a team you want or need, and be able to back that up with some either ways to improve the business or ways to even improve the organization as a whole.
And so that's when I really got the feel for like, once you have a vision, and you can sell that vision, then you have the ability to build as big as a team you want or need, and be able to back that up with some either ways to improve the business or ways to even improve the organization as a whole.
So with that knowledge that I gained through that, through experience of dealing with getting consultants or getting new hires and coming and defining more strategy for that vision, I moved on to the current company I'm with, Gen Re. And it kind of fostered from there that, let me take this vision, let me define a little further the way I feel it should be at this company, and then let me build a full strategy around that. Let me figure out from day one, when you get the data in, all the way to building applications that use that data or the self-service round of other people wanting to use that data for their analytical purposes, let me build that strategy around that.
Defining vision and strategy
So if you take a step back and you think of the data you have, and think of all the data processes going back and forth, how is it actually going to work in a sense of data ingestion? Do you want a platform there, or do you think that it should just land in some folder? Do you feel that everything should be in a database, or do you think something should be in a data lake? So once you get the data into it, now what happens to it? Where does the data processing happen? Does the business units themselves are going to be doing all of that, or should it be centralized?
And then from that point, now you have the data process, now what do you do with it? Do you just create some canned reports and be able to have the executives to see that, or do you go to our stream and tell your business, now take that data and do other stuff with it, create their own report, get their own analytics analysis? Then you got to take a step even further back of, well, what if you're in a global company? Do you have to do this all at a central location, or do you have to split this up into regional locations?
So it's really laying out the whole piece, the whole vision of how everything is going to be. You didn't go into the technicals yet of how to actually accomplish that, just how to view what it all looks like. And that's the first thing you got to sell, and because at least you have a vision, you have something you can communicate to everyone. And then that's when you kind of break down now, what's that strategy to accomplish that vision? Because you can't do all that in one day. It's going to take some time. So you have to have individual steps or a strategy to accomplish that.
Yeah, so it got to where, at a time when I was a little bit earlier, it got to a point when I was leading a team and I was asked to, you know my vision, how are you going to make that vision happen? Once I was given that and when I can sit down and actually really think through that, that's when the moment excites in my head, like this is my next step, this is what I want to be able to do is both to find that vision, to define that high level strategy to do it, and then work with teams to actually make that happen. And that was the time when it clicked on me, but I was excited about it a couple of years back.
Role at Gen Re and data services vs. data science
Yeah, so Gen Re is a general insurance company, we're a global company, we work, we build the insurance policies directly, and I oversee the data department, so I report to the CTO who oversees the whole IT department. My job is to oversee anything regarding data, so we can think of databases, data lakes, data reporting, data science, so anything from IT perspective, I oversee that, I create the processes around there, I oversee the process around there, each of those verticals would have their own managers in there that will take the overall vision or strategy and see it through.
Yeah, so if you think of someone who's going to bring that service up and running, so I think RStudio Pro would be a good example, so one of the things I did when I first started was like, let's put RStudio Pro into our ecosystem, and so let's bring that, let's get all the IT team all together, let's get all the infrastructure in, we had to get all the pieces, so once you get, then you get installed and get it up and running, so that is the self-service or enablement that my area do, so the services we provide is bringing that up and running and providing it for the business to use, the data science from the business side, now use that service, use that platform and build data science related products, might be an office machine learning or whatnot, so it's kind of like, we build it, they will come and use it, and if they want more features or they want changes to it, then they'll work with us to actually make those changes happen.
Career progression advice
The next step is the architect. The architect is someone who now would design that solution, right, so before an engineer would come and solve it, well, an architect might have a high level view of how to solve it, so it might be a hotfix, and you might just do a little bandaid, or it might be a whole architecture change that has to happen, maybe moving on from having the code embedded in there versus having API instead, so it might be a whole architecture change depending on what is the extent of the challenge.
Yeah, so there's, like, two routes you can go. You can go continue that principal route, which you're, like, individual contributor knowledge very much of that. That's also another flavor is a tech lead of that, meaning that you will oversee other people without, like, but you're not directly manager-like, which is, which I did earlier on in my career. I oversaw a lot of different programmers, and I was, well, they weren't on my direct reports. I assigned to work. I did all that, so everything but the HR stuff, and then, of course, you can go to other routes, which you get a little more disconnected from some of the work, but, yeah, you go to people manager and then manager of managers type of role.
Indicators of career advancement opportunities
Yeah, so, when I wanted to, or when someone wants to, like, take more of that management role, then, it depends on the company, but what could happen is you start asking for consultants, or you start bringing consultants and start mentoring them to solve whatever problem it is. And once they start seeing, your managers are starting seeing that you're doing very well with doing the interpersonal relationship and all that, then that can help identify that you're ready. And, of course, communication with your manager is the key. I've always been very close with my managers to make sure that I've known what are my goals, what do I plan to do, what do I want to do in the future, and what can I do to demonstrate that I have that ability, or what are the things I should improve on to show that I'm ready for that or will be ready for that.
But then the other thing is also being very, the communication ability is also key. When I was way in the beginning, I wasn't good at presentations. I wasn't good at speaking in front of a large amount of people. I was very shy, and I can go very technical really fast, and I just didn't know how to slow down and make sure that, or be kind of says that not everyone understands the tech as well. So I had to change the way I'm using the wording. And so once I did a lot of soft skill growth in that, so I did a lot of different classes and whatnot.
And so once you start learning those skill sets, that actually helps you to move on to move on higher because what I find is it's not really the beauty of that coding work and all of that that really the managers see. It's your communication on how to present that and show why, and describe in words on why was this difficult challenge or what benefit does this challenge itself and this challenge give. And it's that that actually gets noticed. It's not, fortunately, not the work behind the scenes. So like, for example, you build a giant application that can do so many different things. They don't see those tens of thousands of lines of code, or hundreds of thousands of lines of code. All they see is the app itself. So then that is your key selling point is what is in the app that's useful there.
Challenges in setting up data science environments
Yes. So the biggest challenge was, at least in the beginning. So a little while back when like analytics and data science was kind of new. And so you're trying to bring in platforms that no one knew how to do or install because it's new. But then, too, is a culture change of allowing people to start having access to stuff and start building products on the business side. And so that was one of the biggest challenges was just changing the culture.
And then in addition to that, when it comes to the technical side, is now it's no longer having an application connect to one thing. You want to have an application that can just in general connect to anything. And so that's also a little shift of how things work. So at the time I was in business, we had to come up with our own test cases to run through all the database connections or run other jobs to make sure things work. And now the only way you could bring it into the company was by doing that, because otherwise then it becomes a full IT owner owned project and all that. And just it's a little different. So nowadays, I think companies are getting or IT departments are getting much more open to these analytics systems. But I think back then it was a little more challenging.
Understanding IT's perspective
So it was a lot eye opening for me when I'm moving from business to IT. And I learned and appreciate a lot more of why it's so difficult. And there's a lot of things we don't think about when we're in the business side. I mean, we do, but how much work it actually takes that we don't think about.
For example, when we say let's bring a platform into RStudio, when I say bring RStudio into the environment, we can just make a request like that and then hopefully one day someone will bring it in or IT will bring it in. But then let's take a step back and see what IT does for that. Well, IT would first need to do a full review of that product. It would open up to security concerns, the data loss. What are the installation tech needs to be? So meaning like do we need VMs? How big of VMs do we need? How big of a use case do people have? What about other use cases that are going to be leveraging this?
For example, whatever you want to use, TensorFlow or, you know, it's a modeling software. So that's an extension to that application. What about communication? Like you have RStudio Connect. So who's going to actually monitor that? Are we going to bring it up and running and then someone else is going to oversee it? What about actual support of these things to keep it up and running? So if a ticket comes, like you can't access RStudio, who's going to support that? So we have to make sure it's a support team actually in line for that.
What is the SLAs? I mean, what is the third level agreement? Meaning that if I'm making requests to have access to the service, well, how long should it take for that to be actually done? And then who actually is going to fulfill those? Is it the data side? Is it the security side? You've got to determine all that. What about the whole architecture? Like if you're a global company, well, you need to think about do I need regional setups or do I need just one giant one? What about all the databases that connect to you? Consider data privacy, data localization. And what about a failover? So meaning that if your whole IT center just turned off, how do you make sure everything comes back on and running? They're running somewhere else. Disaster recovery. All those things need to be thought about. All those things need to be considered. And so the appreciation of that happens when I move over. I get these questions now all the time when I'm putting in a new system.
Impact and culture change
Yeah, so I talked about doing my actuary days. So there's a couple of products we brought up, and one of them was dealing with data validation. And so this is one of the first big applications that we're using Shiny to do. And the accomplishment came from the fact that this was business created, business driven. This is a very different shift from the old days where the IT side would build applications, build everything from scratch. Instead, business built it. And not only did we build it, but we had test cases. We used code. We used databases. We had staging. We had dev staging and prod. We have all those different things that a general IT team would have. We rolled that out and made it available because now it's on the web page and people can access it without needing no code.
I think getting the right personnel in place. So finding people that are passionate about data science or whatever topic it is and being expert in that and both the engineering level, like I was saying earlier, but also at the leadership level. I think that's a way to make things change because at the end of the day, people want to hear things from the experts and people who have had a lot of experience doing that, both from leadership and from actually doing the work.
Staying connected to the work as a leader
So it's like I'm trying to stop myself from jumping into a project and start running code. And it's more because I need to be, I want to be able to more mentor others to be able to do that and to the different wireframes I build or oversee or the methodologies that I've done. I have spent more time on strategy and more of those type of high-level questions and discussions.
I still write R Markdown books internally. I still write little applets or small pieces of code if something's really challenging. But, yeah, I haven't really coded, I want to say, for a couple of years now. I come from the old days. So base R was my background. So I can write base R very, very thoroughly. So it's always fun when some of the new people will come up with a tidyverse and all that and do that. This is the way to do it. It's the only way to do it. And it's a little fun to be like, okay, you can do it this way. Let's do a little benchmarking. And I'll show you some differences you can do. So it's always fun when you have the old timer trying to show the new people there is some stock in the old knowledge.
Managing teams and building trust
As a manager, manager of managers, my job is to make sure that the overall direction the vertical is going continues the right path. So that's basically how far I go with the management of managers. I try to make sure that we have a shared vision or shared idea of where we're going to go the next year, the next five years.
When it comes to the mentoring part, it really depends on what vertical we're focusing on. Because I have a lot of trust and a lot of faith in each of these verticals. And so I let them drive where they go as long as it follows that vision or follows our path. And I let them escalate things to me where things are either got a little bump in the road and need some course correction or we need additional resources or whatnot. So, yeah, this is very much of a trust thing.
Coming from a coding background, or like a technical background, is that I have a flavor of both the way to communicate or to change the language from a technical language into a business language because I've been in business I can talk statistics helps a lot. So having someone to be able to do that translation helps with the communication with business, because if you don't have that things get lost in between. You don't want to get to a point when you're over complicating and with technical knowledge because sometimes business doesn't care about that part they care about what's actually being delivered.
And then from the IT side. I also would need to be able to come up with actual technical requirements or a technical needs that I need to get from the IT side so maybe it's infrastructure maybe some security, which are technically inclined. And so I need to make sure I can also talk in that language. So it's very much of being, being someone who can do both of those communications or interpreter, as well as someone who can have the direction of where they want to go with it all, but how that being able to be interpreted at the same time is, I want to say, very beneficial.
It's more of convincing on the priority of different projects. So, as long as there's alignment, along the senior management of it, these are the projects that are going on, and these are the ones that are getting prioritized and if there's a resource need there then that's kind of like where that discussion is happening. So as long as there's alignment and discussion happening then it's fine.
Advice for first-time managers
I'm having a lot of check ins in the beginning. So, so there's a lot of different approaches that I've tried a couple of them is during a stand up so meaning that having those five minutes, 10 minutes, call for the whole team figure out what all the, all the times is this lovely also have some interactive personal pieces there. But then also having those one on one to like so each of my direct reports, I have a one on one with them every, every week for an hour. And that's just to talk about random stuff.
And it's not really for me it's never been about like asking for permission for stuff it's more of what is, what are you trying to solve. What is your solution. And then does this, does this make sense and let's ask the right questions and then we can come up with a solution together. So it's trying to, again, build that trust with, with each, each of your direct reports so you oversee an engineering team, you want them to try to give you what are the solutions, what are their solutions because you got to weigh it. You don't want to always like say that, oh, this is the only way. There's, there's many ways it's just this different, different approaches that might take different times and different that so that's where you kind of level it and kind of talk it through.
Keeping up with the industry
Yeah, so I, I always look. It depends on whatever I'm working on a time so for a while I got really big into visualization, so I will look up what are the best books about visualizations what are the best authors and I'll buy a bunch of books and read those and then, then the subject may turn into what about leading a data area so the books and so it really depends on what subject is of interest at that time or related to whatever I'm working on. I always, I tend to go the book route that's a hard copy route. But if I'm just like wanting to know what's going on insurance tech I would just just hop on internet just look up some articles or whatnot but in general, it's always books related.
Facilitating team growth
Yeah, so a couple of different ways. So, one is a little bit to my earlier example is that I would like people to come with me, come up to me with the challenge but also the solution at least the initial scratch of the solution, and let's work together. And so, that brings ability to one learn how to communicate a little more to thinking really about the challenge and then defining a possible solution or even just starting at that like what what can we start with and work from there.
And then, and then also the giving them bigger projects like I have projects that range from little tiny ones that might be a couple months work or you have a couple of year one so putting into different projects and and I'm always again always very open. I always want my drug report to tell me what do they want like I always think I've done this maybe like two or three times a year, maybe a little more like what are your goals. What do you want to do, where do you want to, where do you want to be in a couple of years or what kind of challenges you want me to give to you to grow.
And the other side of that is I have a budget for doing seminars or conferences or whatever. So I encourage those to go to those like, for example, one of my direct reports went to our studio conference, because they oversee our studio, so it made sense to be able to go to those. So I'm very much of encouragement to go to those that I wanted to experience that learn from others, not everyone does can know all the different pieces that are so you can do so there's a great opportunity to talk to others about that.
Working with consultants as a leadership development tool
Yeah, no. You want to be more of a, I think, either or if you want to be a tech lead or a manager, you have to know how to work with other people. And, and so, because in order to lead, either people person or as a tech person like you really need to be able to communicate with others, teach others mentor others or whatnot. You could just be a individual contributor who is only focused on I give one given a task I'm going to do that very well and that's perfectly fine. If you want to start working with other people like you need to find an opportunity to be able to be doing that.
Leadership styles
I think it'd be more of a hybrid of two, because if you if you just focus on leading through processes or procedures, then you're more of a technical leader in the sense of I built, we've defined a process and you just follow this. To me that's a little more of a technical leader. But on the other side of that, you need to be able to also communicate and I think that is where it gets you to more of the manager type of leader or the leader manager managers, because you got to communicate outside of your team. So, you might be leading your own team might be doing on that but you can't get from there to outside of your team so other managers other leaders, then you're, you're going to be kind of stuck.
Exciting developments in data and tech
What's exciting and we haven't got a chance to really spend too much time on it but is the idea of the AC interface to build training apps without code, meaning that you can create the UI generator. That's actually something that I was trying to do a couple years back. I would do that. And I got pretty far on it but then things shifted, and then I thought how to stop that project. And then, but I've seen them like wow, this is actually very nice better than what I was doing. But yeah, no, that is, I think, opening a door for new things to happen. And so I'm excited to see what is going to come from that. Yes, but it's something we want to dig into, and I have a feeling it's going to open up more users.
Breaking into tech with no experience
So I think it really depends on what you're doing now. When I first started working, I was on the split team one was focused on modeling and other one was focused on actuarial. So, and they're very split, and since that was only the output of the modeling, so maybe reports and whatnot were actually used by the experts, it was never the code part. So I tried to then, during my work I find some things that may be inefficient or maybe it can be improved on. And so I then brought my knowledge and coding into that space, take all that as a side project and create some process that might be simplified. So I learned everything from the ground up from that, and then not one improved my own skill set, but then to add it demonstrated to management or to other other individuals of here's the potential of using tech.
Tips for data scientists working with IT
What are all the things that you can want to do right so it's not as simple as saying, I want a platform to be available for data scientists and for myself it's more of what are you really trying to accomplish with that platform because it might be modeling is a little bit more complicated than I think just saying that I want a platform. And so if you open those channels of communication, get the support of upper management and it. And of course your own business side. That's what you can accomplish those things.
If you don't have the support of the IT side or imagine on either side, you're going to be pretty fast. So you really need to make sure that communication channels there and if you're not that person that is ready to take that on in the sense of having that communicate open channel on that make sure to find that person because you. There's a lot of communication needs to happen.
If you don't have the support of the IT side or imagine on either side, you're going to be pretty fast. So you really need to make sure that communication channels there and if you're not that person that is ready to take that on in the sense of having that communicate open channel on that make sure to find that person because you. There's a lot of communication needs to happen.
Yeah, LinkedIn is probably the easiest way so I think my LinkedIn is somewhere on the original chat so that's probably the best way to get ahold of me. Emails pretty tough.
As always, thank you everyone for your participation and Matthew. Thank you.
