Today, I had the pleasure of interviewing Hunter Kempf. Hunter is a Data Scientist at Cloudflare, a Z by HP Global Data Science Ambassador, and is finishing up his Master's in Cybersecurity from Georgia Tech. He completed his undergraduate degree in Electrical Engineering from the University of Notre Dame and went on to work at AT&T in a variety of data roles over the course of 4 years while earning his Master's in Data Science. In his free time, Hunter works on a variety of data science-related side projects some of which end up as Medium articles, and enjoys backpacking and hiking. I had the pleasure of working with Hunter on the Unlocked campaign this year and in this episode, we dive into some behind-the-scenes details. We also learn about some of the incredible personal projects Hunter is working on. Hopefully, some of his inspire you to look at your own data in a new light.
Unlocked - hp.com/unlocked
Transcription:
[00:00:00] Hunter: I think the beauty of data science really is a, kind of, is pervasive in every industry, every type of job, or like, if you're working on something, chances are there's data related to it. And there's probably more data related to it. Then there are people working on the data, right? So businesses would always like to have more insights.
So even if it's just, you know, working on something on the side while doing your normal job, I think there's plenty of opportunities around that for kind of personal growth and for showing your value to like wherever you're working.
[00:00:44] Ken: Today, I had the pleasure of interviewing Hunter Kempf. Hunter is a Data Scientist at CloudFare, a Z by HP Global Data Science Ambassador. And he's also finishing up his Master's in Cybersecurity from Georgia Tech. He completed his undergraduate degree in Electrical Engineering from the University of Notre Dame. And he went on to work at AT&T in a variety of data roles over the course of four years, while also learning his Master's in Data Science. In his free time, Hunter works on a variety of data science-related projects, some of which end up on Medium as articles.
And he also enjoys backpacking and hiking. I had the pleasure of working with Hunter on the Unlocked campaign this year for Z by HP. And in this episode, we dive into some of the behind the scenes details. We also learn about some of the incredible personal projects Hunter has been working on, and hopefully some of those will inspire you to look at your own data in a new light.
I always love talking with Hunter. I think you'll really enjoy our interview. Hunter, thank you so much for coming on the Ken's Nearest Neighbors Podcast. We've gotten the chance to work together quite a bit recently with the Unlocked campaign. You also have a really incredible story with your personal projects and your experience with cyber security and soft of creating a hybrid role for yourself there.
So I'm so happy you could come on to talk about both the work that we've done together and your experience in this sort of unique domain within the data umbrella. So again, thank you for coming on. How are you doing?
[00:02:08] Hunter: Doing well. Yeah, thanks again for having me. I think we talked about this for a bit now and it's fun to actually you know, be able to have the discussion sit down. Yeah, it should be an exciting chat here.
[00:02:22] Ken: Heck yes. So the first thing I want to always ask, just to get the listeners familiar with who you are is a little storytelling. So how did you first get interested in data? Was there one pivotal moment that happened that you're like, Wow, I love this. Or was it sort of this slow progression over time?
[00:02:40] Hunter: Yeah, I would say like maybe more towards the slow progression. I got started in data and really like interested in it. More once I exited college and was kind of onto my first like real adult job and, and I had basically done undergrad in engineering and electrical engineering, and then was kind of in a corporate job at AT&T and realized very quickly that a lot of the hard technical skills that I had learned of, you know, the physics and the you know, math side of electrical engineering, I just wasn't using that much, but I still wanted, you know, some technical skill that I could be useful to having.
I think I saw a lot of people that were struggling with data in and around kind of that business side of things, whether it was just, you know, pulling data from SQL or visualizing anything more complicated than just kind of simple bar charts or line charts. And honestly, like one interesting aspect of my journey was AT&T had and still does at this time like a partnership with, I think, two or three different universities with Georgia Tech, Notre Dame, and Oklahoma to basically offer discounted master's degrees for their employees.
And so, you know, for that, it was obviously a relatively easy choice. If you're going to kind of get a discount while working on something that the business is saying, you know, Hey, this is a need, we have that people you know, want to, but like we need help with. And so, yeah, that was, I think, kind of the first real.
Part that got me interested there. I think, you know, that obviously grows and your initial interest in anything, you just get the surface or kind of the like what, what the world thinks about it. You don't really obviously get any of the deep, detailed, like area that you want to go into. I had no idea like how broad data science was when I first started, you kind of jumped into it and you think, okay, it's all like one type of modeling.
And then there's really a bunch of different branches. And you know, that idea of what is the data scientists can mean very different things from company to company or person to person. So yeah, that I think was kind of the early, what, what got me interested. And then obviously we can talk a little later on what has made me stay interested and, and the projects and stuff that I've enjoyed.
[00:05:28] Ken: So, you know, you go, you're working at AT&T, you see this opportunity for continued education and you saw, you know, data is a relevant field and you sort of pursue that, you know, further education. And that has been, that was a little bit of your first exposure to the field. There is where, Yes, you're seeing it in practice, but to get your hands dirty, your first step was in this master's program, correct?
[00:05:53] Hunter: Yeah. I had been doing some stuff at work, mostly with like, I had kind of figured I wanted to learn Tableau or Power BI. And I was able to through work to get kind of the licenses for those. So I've messed around what played with that, but really the first foray into kind of hands-on like modeling anything more than just a linear model was masters for sure.
[00:06:20] Ken: I really liked that because it is practical education where. You had a reason you're at a discount. You like that for me is one of the biggest reasons why I generally don't recommend people pursue master's degrees is because at least in the U.S. This is different globally.
The cost of tuition is fairly expensive. The opportunity cost of time is also relatively high. And with the situation you were in, you were able to mitigate, literally both of those things, you still had a job that you were working. And so the opportunity cost of time going to school while not earning income was, was non-existent.
And the, you know, the actual financial cost was reduced by the company itself. And if someone is curious about programs are getting advanced education that you think that's something you need, I really recommend like pursuing maybe a larger company or one of these... AT&T is not unique in that they have that.
I'm pretty sure like GE where I worked with her a little bit also has a program like that. There's a lot of programs or a lot of companies that value that. And you know, why they value that I think is pretty interesting. It's easier to grow employees internally than to hire people. Right? The cost of hiring a new employee is relatively expensive when you can just have them learn and teach them the exact skills that they need for their role and retain them for a long time is very powerful.
[00:07:51] Hunter: Yeah. I think in general, like if you look at that kind of continuous education master's program, especially around data, there's a lot of different options, I think, unless you're going to go the route of getting a PhD or something along those lines, like doing it while you're working in a master's while you're working, obviously to me was the right fit.
And I think for most people probably is the right fit, going back full time. You're losing out obviously on the income, but you're also, I think most of the practical applications of data science are really in the businesses where if you think of kind of traditional masters, you might have to go to a research university for the, you know, to have the right lab equipment or the right you know, professors to learn under, I think really for most of the data science, kind of the right, you know, equivalent of what lab equipment is, would be large datasets or kind of, you know, business problems that need to be solved with data.
And those are mostly in companies. You know, I think there's very specific problems that universities have you know, advantages around data science, but for the most part it's companies. And then yeah, if you look at kind of the people that you work with, Again, I think most of the people with kind of the best experience are actually out in industry, not in like research university at this point.
So yeah, I think there that pairing that of being able to learn some of the academic side, but still having it rooted in that reality of, you know, a real problem that you're solving with the real constraints of business. Anything you get in the academic setting tends to be a little bit more, you know, the data's already cleaned and ready.
You just build a model and, you know, you report your results. You don't actually ever put the model into production or deal with any of the constraints around that sort of stuff, which oftentimes is a lot more of the job is kind of project management or dealing with constraints of, we have this kind of data.
We need it in this kind of latency. It's gotta be deployed in this kind of environment. Those kinds of things oftentimes can be. A bigger kind of learning curve than just, you know, like doing an SK learn model and fit, you know, predict like, Oh, wow. That was very fast. Like, I feel like a genius. And then, you know, in reality it's like, yeah, that was the short part.
And there's a lot of nuance there, but for the most part, it's kind of the stuff around it that often takes people more time and, and I think is valuable to learn as well. It be hard to learn that stuff just, you know, strictly in an academic setting.
[00:10:37] Ken: I agree. I mean, something during my experience, I did something similar where I did my master's in computer science while I was working full time.
And the coolest thing for me is I was able to take what I was learning and apply it to data at work. I mean, yeah, I role probably had a bit more flexibility and creative freedom than, than most do, but if people know that you're learning a certain skill set in the workplace, They kind of like, if they're your friends or if they're good coworkers, they try to generally try to give you projects that help you to flex those skills.
And I think that that is again, if you have like a really good relationship with your manager or your you're on a strong team, that's something that you can have conversations around, Hey, I'm learning this. I want to work on this in some way. It doesn't always work out, but there's a lot of times where you can control your workflow to work on things and skills that you're actively trying to develop.
[00:11:33] Hunter: For sure. And I think the beauty of data science really is it kind of is pervasive in every industry, every type of job, or like if you're working on something, chances are there's data related to it. And there's probably more data related to it. Then there are people working on the data, right. So businesses would always like to have more insights.
So even if it's, you know, working on something on the side while doing your normal job, I think there's plenty of opportunities around that for kind of personal growth and for showing your value to like wherever you're working.
[00:12:13] Ken: Incredible. Okay. So you're going, and you're, you know, you pursue this master's degree while you're working. How, how does that forward your career or what are the kind of next steps either at AT&T or beyond that?
[00:12:26] Hunter: Sure. So yeah, a little interesting side note, I guess, about the role I started with at AT&T was it was a rotational program with a focus kind of on development. It was called the technical development program and it's still there in a slightly different shape and form than when I was there a couple of years ago.
But in essence, the goal of that program was to kind of give people early on in their career, kind of out of university a chance to try a couple of different roles. So my first role, I ended up kind of. It was with a business unit that was dealing with financial planning. And basically there was enough data there where I was kind of doing the role by the end of my rotation, there of kind of a data analyst taking a lot of the data and creating visuals or reports for people and kind of helping to speed up the decision-making process there by that, by the end of that first role, I was enrolled in the data science masters and was really looking to try to get something in data science or data engineering for my next rotation.
And so I ended up on a role on a team that was building basically an attempt that an internal Kaggle like platform. So you know, kind of hosted internal data sets where teams that didn't have too much data science skills. Could kind of post their data and then basically anyone within the company could view that data and try and like build a model on it.
Usually that data was kind of framed in a way where there was a target column or, you know, variable that you wanted to predict and then kind of enough data around it to like facilitate those. And, and so I actually ended up on that team and was basically working with those business clients that didn't have too much data science skills to format all of their problems into things that would fit into kind of a data science problem.
So if you think of kind of an internal problem, like we want to forecast bandwidth usage or, you know, the number of customers that we'll have in certain areas, all those kinds of things, like tons of different questions would come through. In essence, it was like hearing the problems that they had.
And then trying to say, does this fit into regression classification? What columns would we want to predict? Are, are there variables that we would include in this that we would know, you know, after the fact and if so, we have to, you know, find ways to either clear those out or find other variables that we would know at the time of theoretical prediction.
So that was, I think really cool for me. We did, I think around a hundred projects, like through that pipeline while I was on the team. And it got me a lot of experience just in those kinds of flexing, the muscles of not the modeling aspect, but kind of all of the other pipeline aspects of data where it's, where do we get the data.
How do we format the problem? How do you deal with the kind of those clients or internal stakeholders that might not be all that data savvy, but kind of have the data and, and own the problem? So yeah, went from there and then my last rotation, while I was in the program, I did data engineering on a team that was basically focused on fraud and security use cases for AT&T.
So if you think of any time someone walks into a store there and you know, has transactions, there are models that run that kind of try to predict the likelihood that a transaction is fraudulent or you know, more on the corporate security side when people are logging in on the VPN, how likely is it that they're, you know, not who they say they are and things like that.
So with that, basically those kinds of three rotations, I got to taste. The various flavors that are around data analytics, data science and data engineering. And then when it came time to kind of leave that rotational program, I offboarded from the data engineering role on that fraud and security team to kind of a full-time data science role on that team.
And stayed there for, I think about a year and a half or so, and then ended up at CloudFlare, which is where I work now. And yeah, it's been a, been a bit of a journey, I think throughout that time especially the second half of my time on that front security team, I started realizing. Hey, I'm pretty interested in cybersecurity type stuff, and don't really know too much about it.
Let's get another master's degree. So I'm currently finishing that up at Georgia Tech right now. And yeah, the cyber security master's is quite interesting and kind of in a similar vein has been influenced by the type of work that I was doing that then kind of, I had an, I had an interest in wanting to learn a little more and then jumped into a master's. So definitely not for everyone, but, but has been, has been fun and interesting for me.
[00:18:09] Ken: Well, I mean, that's such an interesting story. I don't think we've had anyone on the show. Who's come from a rotational program, but there's so much challenge breaking into the first data science role. Right.
And there aren't that many rotational programs out there, but they do exist again. I remember there was one at GE or there's obviously one at AT&T. I think there's one at Motorola and a couple of these other big technology kind of... Oh, not as easily, I guess there are technology companies, but they're more traditional tech, more traditional companies than like the huge like Fang or our companies that we think about now.
And it seems like that's a very interesting path to take, to get the, at least the title and the work of data science is because you're working in these, in these companies and going around roles. And you're seeing what you like. You're building experience, you're building relative experience, relevant experience.
I mean, if I recall, and most rotational programs, you at least have some say about the next rotation that you do. And so they're investing all this time to train you and build the skillset that you want. To get the outcomes that you want.
[00:19:19] Hunter: And yeah, I think that that aspect to me is what makes working at some of those big companies quite attractive, especially right out of school.
I know when I was applying like a bunch of the big car companies have them as well. I think if you look at kind of the biggest stablish companies, a lot of them have that. And I think why they do it is because people out of undergrad, you know, might have a lot of skills, but probably don't want to get boxed into the same role that they're going to be in for the next five years.
Especially if you, you know, are kind of unsure of the type of role that you're interested in or the type of work. If you're kind of in a rotational program that at least offers you a chance to try a couple of different things. So I think those will become more popular. But right now, yeah, definitely those bigger, more established companies seem to have them.
[00:20:15] Ken: Yeah. I mean, th there's there's also something that I've said multiple times, it's easier to change roles within a company than to go outside of a company and land a new role, you know, so to go from a rotational program to a data analyst role or to a data scientist role within the same company is going to be significantly easier than starting or finding your first role, trying to wiggle into a data analyst or data scientist role and the jump companies to, to get into a more, you know, whether it is the same or different type of role, like that's just going to be farther.
Right? You, you got to, you got to build your own adventure in a sense in some of these programs. And I really encourage people to explore them more often. I mean, a lot of them, Oh, actually none of them that I know of are purely focused on the data domain. You're sounds like the one at AT&T sounds like it was probably one of the closest, I mean, a lot of them...
[00:21:12] Hunter: And that's my experience with it. I think it was more around like it was technical. So it was people from backgrounds of mechanical computer science, electrical, you know, so engineering type backgrounds. But I would say for the people that I know I took a more data focus bath. I know there were people that took a more like full stack software engineering path.
And then there were others that took, you know, along the lines of trying to become a manager of technical roles. So I think like the one at AT&T was also fairly general, just had enough opportunities where I could kind of tailor the path to be a bit more focused on the aspects that I liked.
So yeah, I think there is like that obviously is a nice thing about those rotational programs, but yeah, definitely not at like a hundred percent focused on it, just that was kind of the path. With the opportunities that were available,
[00:22:15] Ken: I have to say, I really loved your role or you building the internal sort of Kaggle competition type thing. I see how that relevant skillset was very useful in our later conversation about creating challenges for Unlocked a there's no wonder in my mind where you were such an while, you were such an asset on the, on that on that project as well. So I am, I do want to dive into the cybersecurity angle a little bit more.
And so, you know, you pursued it, you're on your second master's degree now. I think I'm the only other person, actually. I know one person with three master's degree and then there's us, us two at at, two and ... But I, you know, I'm interested specifically in what, what appeals to you about that realm and how that intersects with data. And now I guess the work that you're currently doing.
[00:23:09] Hunter: Sure. So yeah. I think early on. So I had kind of a similar feeling with cybersecurity as I was having early on with, with data science. So if you think of my first foray into data science was in a role that was, you know, related to, to kind of data analytics, but it wasn't my main focus.
And I had, I think a lot of opportunities there to see stuff. And similarly, when I was on the role at AT&T kind of for my last rotation in the program, and then my full-time data science role the problem domain was very much focused in and around cybersecurity fraud is very related. Since a lot of the models are kind of adversarial in nature and, and deal with, you know, plenty of cases of, you know, people trying to hack into systems to commit fraud, things like that.
So it was again, kind of taking stuff that was related to the work I was doing at work and saying, Hey, I'm pretty interested in knowing more about it and looking around and trying to say, you know, is there some way that I could learn this on my own? Maybe I think of, you know, structured project and, and class format would be more helpful for the way I learned.
So that's kind of, I gravitated towards a master's there. And yeah, I think in general, it fits well with data science, which is kind of an interesting pairing, but the more you kind of dive into cybersecurity, there's a ton of data in and around those systems. There's, you know, millions of log events.
There's you know, plenty of like binaries from suspicious files that people will upload to various sites and, and try to do automated detection of, is it a virus or not? Or what virus is it. And then kind of on the side of the models that anyone builds in cybersecurity, similar to what I was doing, I'm kind of on that fraud and security team.
They're very adversarial. So you can create a perfect model for how the actors are working against you at this moment in time. But next week, you know, they're going to see, Hey, my virus wasn't working or Hey, my, you know, fraud scheme, wasn't working and then they change it up. Right. So I think that to me is quite interesting understanding kind of the parameters, a little better of, you know, from a data side, what are the aspects that people can change about stuff they do?
What are things that are harder to change or maybe impossible. So that was kind of my early on interests. You know, Hey, I'm working in and around this data, I'd love to know more about it. And I was interested enough to kind of dive into a master's. It's definitely a bit more than probably most people would do, but yeah, I've been interested in it kind of this, this whole time and I'm in my last semester of it right now. So, was interested enough to kind of see it to completion.
[00:26:23] Ken: That adversarial nature to me is quite fascinating. So in data or in software engineering, the problems you're working on, they're not meant to trick you. And they're not easy necessarily. You're trying to find insights, but they're not actively like trying to fool you right.
[00:26:43] Hunter: On the other hand, most of the time, they're not there, there are those kind of small few cases, but yeah, for the most part, I would say maybe things like, if you look at social media or social networks, there's plenty of people that try to game algorithms there, but you know, for the most part, people aren't trying to gain kind of any code that most people are writing.
[00:27:07] Ken: That is true, I guess like in financial markets too, it's like, Hey, you know, we're trying to understand what algorithms to like whales are using to like decode that. And, you know, but you know, not that invest based on those, but, but yeah, but I think, you know, the vast majority is not aimed at like tricking someone or like, or being tricked.
And the nature of cybersecurity is quite the opposite. The bad party is actively trying to fool your systems or whatever it is. And that presents really unique and kind of fun challenges, right. Especially for someone who's, who's quite a thinker. Whereas, you know, me, I'm like one of the things I like about data science is that I know like it's always like cathartic to know that the challenges that I'm working on, they're not meant to be challenging, right?
Like it's only me. That's making them more complex than they need to be. Whereas on the other hand, I think it's kind of fun and exciting. Also terrifying to me that the other person or there's other actors that are actively trying to fool me and like that sort of cat and mouse game is also one of the big challenges. Right. So, you know ...
[00:28:21] Hunter: Yeah. I think if you look at traditional like data science or machine learning, oftentimes let's say you're trying to forecast like the number of units or company is going to sell this quarter or something like that that has, you know, effects of other people, but there's no one maliciously trying to buy up a hundred of your, whatever item you're selling.
And then, then that just go, yeah. Like people just don't care enough to do that. Or the resources required wouldn't really give them any benefit. Right. So I think for things like that, there's plenty of domains within data science that, and I would say probably most that you're kind of just trying to solve the, kind of the same static goal of matching, you know, whatever the case is.
And a good model from a year ago will probably fit a good model today. You know, assuming that like the patterns and behaviors are relatively, you know, inland. Where, yeah, a good model yet, like a year ago probably doesn't fit the average attack cybersecurity related today because you know, as soon as people figure it out, the software comes in and blocks it and then it's kind of onto the next thing.
So I think with a lot of that, it makes it interesting. It makes it exciting, but it also makes it, you know, often a problem that you can never really solve. It's a problem that you can mitigate as much as you can for the time being, and just understand that eventually there'll be different patterns or things that slipped through.
And so trying to not only have the models of block things, but also have models that alert on suspicious behaviors or understanding it, you know, kind of okay. You know, once people get into our system, what are they trying to do? So the concept is a little bit of this idea of. Like a zero trust system or a defensive depth people call it, you know, kind of different stuff.
But in essence, like don't just have a big front door with a bunch of locks on it. Have, you know, other systems inside like the proverbial house or, you know, business there that tell you, you know, when something is going wrong, so you can catch it because these models will never be perfect front doors, they can be great, but they, you know, will as behaviors change, like some stuff can get through. So just being able to catch that and stop it early is also, you know, an interesting challenge that ton of data and data scientists work on
[00:30:52] Ken: This episode of Ken's Nearest Neighbors is brought to you by Z by HP. HP's high compute, workstation-grade, a lot of products and solutions. Z is specifically made for high-performance data science solutions.
And I personally use the Z Book Studio and the Z4 Workstation. I really love that the Z workstations can come standard with Linux and they can be configured with the data science software stack. With the stack, you can get right into doing the data science work on day 1 without the overhead of having to completely reconfigure your new machine.
Now, back to our show, I really liked that analogy. I'd never heard that one before. I also hadn't paid too much attention to cyber security. So it might be a really common one. But you know, you'd mentioned before, and it's hard to jump around a little bit, but it just came to my mind in terms of your education and that learning in a more formal setting might work very well for you.
I found something very similar as I like a classroom setting. I like having a lot of structure around it. I know from our offline conversations that you come from a fairly academic family, does that have something to do with, with your, your appetite for structured learning, or is that something that you've developed yourself over?
[00:32:07] Hunter: Yeah, I'm not too sure. I know, you know, everyone learns better, like different ways. I think for me, a lot of what that structured learning and I've done some nano degrees and some of those kind of like just watching some person on YouTube explaining stuff. I think the nice thing for me about any of the kind of more structured learnings is you kind of make a commitment and you're like, I'm actually going to learn this thing.
If you're just watching the YouTube videos, there's no formal commitment. If you're just doing an editor's rate, you can stop doing it at any time. When you kind of like, say, I'm going to learn this and I'm going to do a masters, you start, you know, putting the money forward and there's assignments and you have to do them.
And there's, you know, it's kind of not at your own pace. I think for me, I learn best in that format of kind of. A little bit of like over committing myself and then having to like live up to that over-commitment kind of. So, you know, if I sign up for something I'm going to go through with it and, you know, spend the time to learn it and figure it out.
I think, you know, everyone's different, right? There's plenty of people that can, you know, stick to that learning path, you know, a hundred percent on their own and figure out, I think the other thing with masters, which is kind of nice is, you know, someone else has figured out a bit of that path of the classes that you're taking.
So what fits together, what's relevant towards like basic understanding. I think when you, when you look at data science as a whole, there's kind of plenty of people that say just jump right to machine learning. And I think, you know, you do definitely miss out a fair amount by. If you're just coming up with what you think is the most relevant at that point, you miss out on a lot of the more general stuff that often, you know, no matter the domain is more helpful.
So if you, you know, if you just dive right into machine learning or NLP, or, you know, some various, you know, image recognition, some very specific domain, you're often not going to really understand the context that stuff's in. You might be able to copy and paste the code and get stuff to work. But if someone's asking you to kind of change something in a subtle way, or, you know, tune parameters differently or work into, you know, I think things start to fall apart when you learn that kind of, let's say a little more brittle way of just learning examples, as opposed to, you know, some of the underlying mathematical frameworks and like just general problem solving.
And some of those things, which if you look at, you know, university classes, oftentimes they schedule stuff where you build up from kind of the base. And then once you have the strong foundations of some of the mathematical comp components, some of the statistical components, some of the, you know, basic linear models or things like that, then you start to graduate onto like the more complex stuff and you actually understand the format and, and why things are done as opposed to just, you know, anyone can copy and paste stuff.
And it may work or it may not work, but you're probably not going to be able to kind of troubleshoot as well if you're just copying and pasting. So yeah, I think that that's kind of a, probably a longer winded response than you wanted, but yeah, I think I do learn better in that classroom setting for kind of that variety of reasons.
[00:35:44] Ken: Well, it was not, it was the perfect amount of time because it gave me a little time to think about what you were saying. And one thing that I noticed you didn't really mention was like the quality of the resources, the like the actual concepts that are being taught to me, that's a really important thing is like, you don't pursue graduate education because you can't learn because you can only learn certain topics there. Right?
Like all these topics are pretty pervasive on the internet, but the systems that are in place, if they work for you, that can be really compelling. So for me, I liked getting grades is something that I guess a lot of people don't know about me is I'm like ruthlessly competitive in the academic setting.
Like I like breaking the curve. Like that was a lot of people hated me for that. But on the other side of that, I helped everyone study. So couldn't but, but the idea is that. if the mechanisms in place for accountability and for competition or whatever really motivates you work well in the academic setting.
Great. If not, if you can, like, the biggest thing for me is historically this isn't the case now, but I had trouble like with the meta analysis of data science. So like what should I learn in what order? Right. Universities do that? I think relatively well, I mean, there's some things I would adjust around a little bit.
Like honestly, I would probably put like some of the math classes in the, like, closer to the middle of the curriculum, because I believe in like, okay, if I think we should apply and use that framework of application to understand the mathematical fundamentals, like it'll make more sense after you do it one, but I digress. But the idea is that now, if I were to go back, I could just look at those curriculum. And build my own curriculum based on that sequential order, right? Like...
[00:37:39] Hunter: Yeah. I think a lot of that stuff is like you, if you have some framework of understanding what to learn, then it is different. Right.
There's and I think there probably are good resources out there of people that have kind of come up with the topics or stitch together, some of the like various resources and, you know, have like maybe communities on discord or on other places where people can kind of get a little bit of that. Like, Hey, we're learning together feeling or, you know, kind of an ability to work together on.
I think, you know, with, with all that, then you could kind of simulate the same education that you get from you or see, you know, without the grades, maybe you wouldn't be as enticed, but I think, yeah, it's hard to stitch those things together, especially if you don't have much experience in the field.
So yeah, I think for a lot of people you know, if you find that right resource or if you're not sure that you want to fully commit to something, right. Cause you know, a master's degree is not a one month thing. It's not a it's for the most part, like a two year commitment kind of and I, there are shorter ones that are in person, but if you're going to do them online, you know, the fastest, you normally will do it as kind of around two years. Right. So definitely a big commitment there. And I think you know, for some people it works better than others, obviously, you know?
[00:39:11] Ken: I mean, there's also that other angle where like, In a master's degree in college, even like you're paying for the course and you could go and fail the course and not get credit for it, which is fascinating.
Right? It's like, Hey, I'm paying money for this. And you're going to fail me with the hat. Like I never thought about how strange that actually sounds like when you think about it. But I mean, in a lot of certificates, that's probably not, that's not like super relevant, right. Where it's like, Hey, I'm paying for this.
I just have to go through the motions. And I complete it. And I got the certificate at the end. Like there, the downside risk is not as much. And, you know, I will say something. I like that. You know, a company I'm affiliated with - 365 Data Science has done is they've gone and like really put some rigor into the exams and the certifications that they're doing.
And like you can actually fail, which, you know, from a pure profit perspective might not be the most effective on their part. But on the other side, from like an actual learning perspective, I think it does. Add that additional layer of credibility and like, you know, it's like, Hey, I have to take this seriously.
I have to like dig in and want to learn. And if I can, and, and you know, that's probably why a lot of people hate school, but also at least for me, why I liked school is that there was that silver double-edged thing is especially on the accountability form.
[00:40:32] Hunter: Yeah, sure.
[00:40:34] Ken: So something that I really like about, about you that we've talked about offline is that you just like working with data and you look exploring it and it's not there isn't, it's not like a means to an end. It's well, I guess in some sense, that is, it's not a means to like a portfolio or like for landing a job or something like that. You're doing it because a lot of these projects are relevant in your life and. To like help understand you better.
And I personally resonate with that. I don't actually get as intense into it as you do, but can you talk to me about some of the projects that you've done that have been just like relevant into improving your, your everyday sort of livelihood?
[00:41:18] Hunter: Sure. Yeah. I've delved quite a bit into this sort of realm of personal data collection and trying to understand, and like use the data resources that I have learned from classes or, you know, jobs I've worked in and trying to say, Okay, how do I apply those?
I think to me, and this is different for everyone, but to me, like the most interesting data I could possibly work with is data about my life. Right. Because it's the most relevant of anything that I could have. So yeah, I've done it. A couple of different things. I guess around, if we look at kind of early pandemic I was watching, I think like a lot of people staying at home, watching a lot of stuff on like YouTube or Twitch or, you know, Netflix, HBO, all that sort of stuff.
I did some interesting analysis on some of those like shows and movies, trying to understand a bit of those. I did then a couple of things with Twitch streams and like tagging live events that happen in games. So if you think of early pandemic the game among us was quite popular, has a lot of visual cues and.
Within that kind of game context. I was basically training computer vision models that like try to pick out, you know, this player got this many kills or got, you know, this many reports or, you know, the various events in any of those games did this many tasks. And the problem statement, there was just something where I was like watching along and interested, stuck at home and, and you know, figured out like, Oh, here would be an interesting kind of data science problem you know, along that sort of stuff, then plenty of stuff that I've been doing lately around trying to understand my sleep better.
I have like a sleep tracker. That's this big pad that kind of sits under my bed. And every night when I go to sleep monitors kind of heart rate and, and some other stuff there that, you know, tells you, okay, you have. This much sleep. It took you this long to go to bed, took you this long to wake up a bunch of interesting stuff there.
And then you're able to pull that out and kind of compare it to other data like you know, time that I was sitting at my desk, I've done some some work there with like some home automation software and some like off the shelf models for person detection, things like that, where I now have a kind of live look of, I was sitting at my desk working from home this long, you know, today or working on projects and all that long and, and understanding, Hey, it's time to go outside and like walk around for a bit or, you know, do something outdoors, athletic.
There's a ton of different things that I collect. In the last, I guess, few months, I've also for long drives. I take in my car. I've been bringing like a GoPro and putting that on, like facing out the front. Cause I have some ideas on like capturing, like the billboards and media that I see trying to understand like some stuff around that, or just like play around with the data.
Like, you know, do you know how many different types of cars that I pass by or did I pass by the same car multiple times on like a long road trip where it's kind of bypass them. They pass me, stuff like that. Most of the stuff I have around personal data is more, I'll see something or I'll be like watching something and they thinking like, it'd be like some interesting data project that I could do around this.
Or I wonder if someone has some resources for this, like for the dust detector, there's a ton of like home automation and home security kind of open source stuff out there where people. You know, have figured out, okay, you can take this kind of off the shelf, hardware, stick it there and, you know, do your own like private home security.
And then there's, you know, additionally you know, machine learning models that then you can kind of hook into and they do person detection. So building those components, I always have fun just with like little side projects that have nothing to do with work or school or anything else. Just, and, you know, not stuff that I'm like out there, you know, while I want to be famous for doing this, or like put a ton of effort to like, make it like a premier project.
It's just more like small little things that I think, Hey, this would be kind of interesting and, and, you know, put some effort towards it. What ends up happening is a bunch of them are kind of half-baked where they're. But 90% done, but not like the last 10% of automating some steps or, you know, doing the full analysis for certain things.
But that's, you know, part of the fun of not having, you know, strict deadlines for any of this stuff, it's more just, you know, I do it as long as I'm interested, if I kind of lose interest in one realm, then you know, I'm often and there's kind of no harm, no foul there.
[00:46:43] Ken: Yeah. You know, I really liked that you actually at least get to the 90 or 80 or 90% stage.
My problem is I'm always thinking about them and I'm writing them down and I just never start them. So that's one step further. But, but it's also cool that you're constantly like your brain's always going, thinking about these problems thinking is how about how data is relevant? Like, I don't think that that's a prerequisite to become a data scientist.
Like some people can go into the office, turn it on, go home and turn it off and not think about it. But I find more often than not most people who. Work in the data domain are constantly asking questions or thinking about what is the relevant data here. I think about how could data solve this problem or like, is this even a problem?
And like, I find it really fun and exciting to talk to those types of people, because you just view the world a little bit in terms of like problems and solutions. Right. And that's in my mind, a very like exciting way to look at the world is like, Hey, there's so many opportunities where data could do good.
Also a lot of where data can, can cause harm, I guess. And you know that when the world is a series of problems and solutions that are, that are relevant to data, it's not all data. Like, at least for me, that's like inspiring because a lot of people view the world is like, Hey, it's static. It's not something that we can change.
Like there's nothing I can do about X, Y, Z. And with the tools we have today, there's like something you could do about almost. And it'll require a lot of work, but that's, I dunno, that's like the most optimistic viewpoint you can add is like, Hey, anything out there? I feel like there's something I could with the tools I've created it in the data domain and with coding or whatever it is. There's probably something I could at least do to influence it.
[00:48:34] Hunter: Yeah. I think that that's a little bit of that, like almost engineering mindset of in engineering, you learn math, you learn physics, you learn chemistry, you learn all these, you know, kind of building blocks. And in data science, you learn, you know, a lot of the different building blocks of how you work as a data scientist, but the concept of how you take things that you've learned and apply them in different situations, you have slightly different or wildly different.
I think is like part of what makes any of the kind of engineering or. Let's say coding related fields quite interesting is you can kind of take stuff that you would do for a business and you can apply it to, you know, your personal life or, you know, to, to anything else. You can apply it to like the community events you're interested in, or, you know, whatever, like outside of work, that is interesting.
It's, you know, when you have the right tools and you have that mindset of wanting to solve problems, you can do a lot of things, which I think is really cool. And that that's, you know, oftentimes there are people out there that are very committed on kind of singular projects and they make it, their life's work.
Life's work. You can basically kind of come along and, you know, leach off some of it, like, like I said, some of the home automation stuff, like I haven't really contributed too much back to that community. I just basically took a bunch of piece parts that, you know, a lot of people had kind of spent I'm sure, you know, countless hours working on.
And then, you know, I basically take a bunch of those piece parts, stitch them together and solve the, you know, like small problem I had of how long am I sitting at a desk when I'm working from home? And in those cases, you know, I think there's plenty of stuff I can get out of it. I think then the concept of also trying to do things that support back those communities or some communities that you know, you kind of can further.
And I think, you know, with, within that open source community, there's a ton of people in data science and software engineering. Really put a lot of unpaid time and effort to just kind of like air quotes, make the world a better place, which I think is awesome. And you know, that's something, when I look at like any of the projects I do, where I want to eventually kind of like go a little bit more towards, of like being able to produce stuff that, you know, gives back to, to others and is building blocks that other people can use.
[00:51:13] Ken: You know, it's funny, one of my friends Stefanie Molin, who I've had on the podcast before I was talking to her the other day, and she was talking about how she just made a country contribution to pandas and it's because she was having an issue with one of the methods that she was using and it, and it's like, Hey, this isn't working for some reason, let me go fix it.
And she just wanted to fix that. And she's like a contributor to the community and it's like, well, By using a lot of these tools. If you're using them a lot, you kind of see the problems with them or you see ways that they could be improved and that's an opportunity to contribute. It doesn't have to be this huge premeditated thing of like, I'm going to figure out a way to contribute.
That can happen organically. The more you use a lot of these things. And like, it's kind of cool if you're coming in from like a more data background. Right. And you're looking at maybe one of these home security automation, things that isn't super data-related, you bring that data perspective that could potentially add to an open source project.
I kind of view reading in the same way that you view, like going through and looking at these open-source projects. Like these people they spend it's their life's work to produce some of the books that are out there. Right. And I get to sit down and like, get all the benefit from it and like four hours right there. How cool is that? That like, I can. Like 90% of what this expert knows about this certain domain that they've dedicated their life to in like a day.
[00:52:42] Hunter: Yeah. It is quite interesting there of, I think the concept is kind of like you stand on the shoulders of the people before you or whatever, and that's what makes you a lot of our, the world we live in and, you know, especially data science and computer science really accelerate fast is the fact that you can kind of understand too, you know, maybe not the full degree, but to enough degree to actually do stuff with, with the tools people have built kind of even what that expert.
[00:53:16] Ken: Awesome. Well, so I definitely want to make sure we save time to talk about you know, you, or both of us being seen by HP Global Ambassadors, as well as the Unlocked campaign. So that's been a big part of our lives for the last about five months. And like, you know, we could probably do like four full podcasts episodes on that experience, but I definitely want to carve a big chunk out of time here now to just talk about what that was like and you know, just bring people into what the process was for both of these here.
So to give a little background, you were actually involved in the unlock project before I was, and I'd love to hear about how it originated and what were some of the key takeaways associated with early stages.
[00:54:05] Hunter: Sure. Yeah. So I think I joined on. Like relatively early in the process they had a couple of ideas and for people that don't fully maybe haven't watched the Unlocked video or don't understand in general, what, what the final product that that was kind of created is this hybrid ad kind of storytelling paired with data science.
So in essence HP spent the money to kind of build this story that focuses on data scientists, solving, you know problems related to like a story. And then there's four different data components where you basically can step into the shoes of the Like actors or the protagonists yeah. And try to solve those data problems.
And so, yeah, early on, I think they had an idea basically of, they just wanted to do an ad. And, and so there were a couple of people that had some ideas of trying to make it a little more entertaining, a little more interactive. And then that was kind of when I was brought in terms of helping them kind of understand a bit more about the data side of things and understand a bit of kind of, you know, how could we make something that was more interactive in terms of audience participation or or something along there.
So like the, one of the first steps and this, I guess the early on like idea of what, what the general story we're going to do was kind of. It took a while to figure out. So there were a couple of iterations, but the earliest on idea was some kind of like almost murder mystery or something there of like surrounding this idea of a museum.
And we were going to do the challenge about, you know, card swiping and trying to find a culprit based on, you know, someone who manipulated data. And you could figure that out if you looked at patterns and things we ended up changing, you know, a bunch of this stuff, but early on, I did some kind of proof of concepts like data set building of, you know, if we were going to just, you know, have fake data that we could use to solve this problem you can kind of do anything. Right.
And so within that was, was definitely pretty creative and And then got to work with that creative side of the kind of advertise advertising team and creative teams to kind of blend a little of where they thought the story could go and do like a proof of concept kind of first challenge.
And so yeah, I worked with them, did that. And then I think right after like HP kind of saw it and said or like the leaders there saw it and said, okay, this has some legs. Like we could do something cool with this. Then that's about, I think when they brought you in Ken to kind of help also with some of this like ideation of how the story is going to fit, how the challenges are going to work.
I think, you know, obviously having more voices in the room helps to figure out stuff that works, you know, for different audience types. And and it was definitely, you know, an interesting problem where most of the time for data science. Thinking of, I have this static data, you know, I need to solve this somewhat static problem.
There's relatively, you know, kind of rinse, wash, repeat kind of things that you can look and say, you know, this is a classification problem. I have this data. Maybe I could get some more data, but like generally I'm not going to change by classification problem to like a, you know audio detection problem or an image classification, or like you, you know, you're pretty much set in the, in the genre, but this was fun because, you know, kind of, as much of the story was dependent on the data side as the data was dependent on the story.
So, you know, a bit of that kind of play back and forth and being able to really change, you know, from where it was of this kind of museum murder mystery to where it ended up of, you know, a story about kind of an ancient flower and, you know, finding a cure to some rare disease, quite different, but actually like a lot of the core data concepts of the types of problems that we tried to fit in there were kind of relatively consistent. Just the, you know, characters and obviously, you know, how it related to the story changed a little.
[00:59:08] Ken: Yeah. I mean, I honestly, I loved the sort of the museum concept. I mean, we had talked about how cool it would be to have like a generative art type of competition or a challenge on and not a competition.
I think that would have been so fun. And you know, who knows maybe if there's a, there's a next one, that that could be something we bring in. The other idea that I thought was so cool is being on the other side of like, what if. Be the, both the protagonist and like the malicious character using data, whether it's like fake generating articles or something like that.
I mean, we don't want to put that out into the world too much, but it's kind of fun to be able to role play. And that's one of the beautiful things you have about like a fictional piece like this, where there is some fantasy and you've got to work on data that like might not exist. And so I think that that was something that was, that was new for me, as well as like the feature space or the problem space.
Just as you mentioned, it gets so big when you're not only trying to determine like the questions and what data should be there, but also like you have complete free range to make up the problem. Right. Yeah. And, and how often are data scientists in that position? Usually we're constrained by the data that we have, right.
Or we're constrained by the domain that we're in or whatever it is. And like, even you think about how, how, how overwhelming it can be to, to come up with a challenge from complete scratch. Although it was intimidating, I would hope that we did a pretty good job then. And we had a lot of, a lot of fun with it.
I mean, there's something so neat about it. Just like the little, the little nuance that we were able to add into to the challenges. So, so obviously there's four challenges. So the first is data visualization. The second is natural language processing. The third is audio signal processing and the last is some computer vision. Can you kind of talk about how we sort of chose those challenges and the like, and, and also which one was your favorite?
[01:01:21] Hunter: Sure. I think you know, when you look at that big problem space of, okay. We want to have a data challenge that associates with this story, you know you want to try to, at least the way we approached it was to try to give ourselves some parameters to work within.
Obviously we had the creative team that had kind of created the outline of what the story might be. It had kind of, you know, flushed out some of here's the type of characters or here's the type of maybe problems that they're working on and things like that. And then from our perspective, like we wanted to at least have a little bit of a framework of, okay, what kind of challenges can we add in that are, you know, giving a decent breadth in terms of the types of problems that a data scientists might encounter?
So not having everything just be, you know, tabular machine learning. Here's a bunch of columns with a bunch of data, you know predict, you know, classification for predict progression. Like four times we were looking for a little bit of the kind of, you know, journey from, let's say more freedom to work on stuff and more creativity around visualizations and around kind of some of the natural language processing stuff there all the way to kind of some of the more well-defined and maybe more like unique things that people haven't done before around audio signal processing.
And then, you know, again, looking at kind of a classic data science problem at this point now with computer vision or kind of, you know, that classification problem there with, with pictures We did try to have that kind of depth, a little of trying to have different types of challenges.
So each one is a little fresh while trying to not make it too difficult and not make it too easy. So, so I think like, hopefully we did a good job there and, and I think we did, but yeah, like give the audience a little understanding there of some of the parameters that we kind of gave ourselves going into it.
I think my favorite is one that I spent like a fair amount of time on the I guess the tutorial content that you recorded that third one, which is the audio signal processing. We have, I think an interesting challenge there that, is it a twist on even, you know, any of the category challenges out there around audio data of kind of taking it another step.
We have a bird calls that you're trying to classify. This is, you know, has this bird color, it doesn't have it, but then going a step further to say, can we try to count the bird calls that happen in a certain length of a clip and you know, offers quite a bit more nuanced than just build a machine learning model of you have to build a model and then, you know, have some sort of filter code to say, okay, this was one bird call.
This was two, this was three know, kind of do that counting there. And I think it should, it should be fun, should be more than just a you know, kind of a single. Like point and click, you know, here's the model built a hundred percent accurate, everything. Even if you do have a hundred percent accurate model that you have to do a little kind of software engineering or a little like, you know, creating code that that's not a hundred percent ML, which I think, you know, adds a bit to like the richness of that problem makes it you know, a little bit air quotes, more real world.
Can we try to keep you know, the problems somewhat simple, but also, you know, add elements that we as data scientists, if we had the same data might think to actually solve the problems with so yeah, I had a lot of fun working on it. I think really all the problems are, are, are fun, but that third one to me was something that I didn't have too much experience. Kind of the audio signal processing and learned quite a bit while I was actually, you know, kind of working on that as well.
[01:05:48] Ken: You know, it's funny. So I can take absolutely zero credit for any of the written to the like Coda tutorial portion. Hunter did all of that, but I also got to learn a lot, which is a cool thing for me is like when I did the tutorial, I had to at least understand some of it.
So I had to go through and I say, why are we doing this? Why are we doing this? What are the implications here? And I sort of Unlocked a new learning thing for me is that if I try to teach something that I don't know that well, man, there's a lot of pressure, especially if I'm doing it on like a public forum like that, it's like, Oh my goodness, I have to go through.
So if there's any any stumbles that are in the recorded tutorial, that's because I did not understand. The the code well enough, if there's any issues in the code that's on Hunter. So just...
[01:06:39] Hunter: Yeah, yeah. And then I think we did a pretty good division there of the tutorials. We brought in a couple of the other ambassadors and then like you worked on the tutorial for number four and, and, you know, probably have maybe let's say better video tutorial, you know, covering that because of it.
But yeah, I think the other interesting thing kind of, I thought of as, as you brought that up, was this concept of like, when you actually have to teach something to someone, you actually have to understand it all of a sudden, a deeper level than you understand it, just to use it. So, you know, there's plenty of stuff where you can kind of within data science, like build a working model, but oftentimes when you're, when you're at in a job or, you know, dealing with business clients, they'll ask you, how is this.
Not just does it work. It's like, they want to understand a little bit more of like, why does it work? And, and I think, you know, that kind of is a skill that plenty of us have to use a kind of, you know, it's not just enough to have something that works. It also, you oftentimes need to understand whatever the problem or the like code that you're using well, enough to kind of explain to others or, you know, at a high enough level, explain to others, like generally how stuff is working there and, and why certain decisions were made.
So, yeah, I think for everyone in your audience definitely worth a watch of the video. And then the video I think is quite interesting, but the challenges really are where, you know, if you're kind of interested in that data science domain, there's a ton of kind of fun to be had there. You don't have to fully do everything, but maybe choose one or two of the challenges that speak to you and are more interesting and, and give it a go.
I think both of us are watching on the kind of results there to, to see, you know, anyone that posts their solutions to any of those things. Like we're, we're looking in and you know, hoping to see some unique take somethings or people that are just learning from it. So definitely a fun project there for anyone out there listening.
[01:09:01] Ken: Absolutely. So also a special shout out to Andrada Olteanu and Nick Wan, who are the other Ambassadors that help with the creation of the tutorial content? I will say that I had to learn the hard way that you have to learn and understand these things arguably better than if you were just implementing them.
When I, when I went through, did the tutorials, I mean, that's something that you know, it was, it was fun and it was exciting, especially having all the film crews there, but it was pretty nerve wracking, like even a sitting recording for basically like 30 minute hour long tutorials covering tactical things all in all in one chunk, but very different from, from sitting at home.
And it's an interesting experience. I think, you know, one of the nice things is that I do a lot of recording and, you know, just, just having anyone come in and do that probably would have been, would have been a bit much to start out with, but I have to say it was, it was an awesome experience and the interactive nature of these things and the, at least the semblance of education that people can take away, I think is something really unique, but I don't think it's necessarily something that is like unheard of.
I kind of think it's like the future of a lot of content. I mean, you, you had mentioned this offline that, you know, with a lot of the shows on HBO, they do an after show that's content and engagement and you know, where does, where does this come from? Is this the future of content? Is this the future of advertising? Like what do you think about that?
[01:10:51] Hunter: Yeah, I think it's an interesting question. I think. In terms of content, everyone is in this kind of race to make their content, the most engaging you look at, you know, Netflix, HBO, any of those, like companies, either publish numbers about viewer engagement or, you know, internally obviously track that with data science teams.
And so I think, you know, anything that you add that, you know, can be, let's say viewed, consumed, interacted with outside of just the show that these companies spend millions of dollars per episode on is obviously to the benefit of that company, because. If you look at a show like game of Thrones or something like that, you know, there's plenty of rabid fans that are, you know, watching each episode and extremely excited.
If you can think back to, I think it was, you know, 2015 or 2016, there were plenty of watch party type things that you do with friends when the new episodes come out. And I think those experiences show that, you know, for the right content and for the right situations, there's people that want to, to interact and, and kind of do more than just wait for the next season to come out in a year.
And so then people come up with, you know, all these arts and crafts projects around things or anything. And I think it's to the company's benefits to add in as much Fuel to that kind of creative fire for their audience. So yeah, adding additional content out there obviously shows that are based on books have that kind of already, but, you know, for things that aren't then having, you know, additional content that people can consume or interact with or deal with, I think definitely is.
Yeah, I think, you know, around this, this concept of specifically in the data science domain, it's, you know, just scratching the surface, like in, you know, what, what makes education more interesting, you know, could definitely be, you know, something along this lines of kind of, instead of listening to a professor, professor lecture, you kind of get the elements out of a story and then apply it.
Or, you know, maybe you watch the story, see how it could be framed in a problem and then watch a lecture about it or something. I think any of that sort of. Is definitely a cool, you know, future of the kind of industry of learning as well as, you know, content in general. Anything that you kind of have more interactive audience participation will always be better than the kind of show where you watch it forget about it and don't talk to your friends or relatives or anyone else about it.
[01:13:45] Ken: Yeah. I mean, I would love to see some of this stuff be used in like a classroom setting. That would be a really fun type of type of course or does. I mean maybe selfishly speaking, I think it would be fun, but you know, there's also, you had mentioned, this is just scratching the surface. There are so many things you can do even to make things like this more engaging.
And, you know, we, we talked pretty extensively about even breaking the fourth wall further and having like websites that were built out or Reddit forums that were built out for, from the stuff. And actually using live website data or you know, building an API that could be used to classify images after you built it right.
When people could upload pictures of their own flower and see if it's close to, to lodge or not, or something along those lines. Yeah. I mean, the exciting thing, I think, especially for me, is that the upside, the potential for future projects or anything along those lines is Islam, Ellis. There's such cool stuff that we could do.
And, you know, I think we're obviously both hoping this is successful enough, that it would breed additional really full content.
[01:14:56] Hunter: Yeah, for sure. And I think like with anything it's, if you don't believe in the like, project that you're working on, you're not gonna put forth full effort. I think both of us were interested in and thought this had, you know, such good potential and think.
You know, worth doing right to where we put in a lot of that time and effort towards hopefully making this a success. And then, yeah, hopefully there's, there's other projects that either try to take some of the components of this or, you know, hopefully HP does it another one next year or sometime in the future.
[01:15:35] Ken: Heck yeah. So also a special shout out. We want to make sure we think Natalia, who was who worked with us as well. You know, she, she had some other work obligations come up, but she also made a, made a meaningful contribution. You know, those are all the questions I had Hunter. What do you have going on in your life? How can people connect with you or learn more about you for anything related to that?
[01:15:59] Hunter: So yeah, I guess it's an interesting thing there. I have a LinkedIn and I don't have YouTube or anything there. I occasionally will write medium articles on stuff I'm interested. We worked a little bit together on kind of some March madness stuff, which might be top of mind as people.
And I have a week or on my LinkedIn links to, to the medium article, kind of the high level steps that we went through towards, you know, trying to generate features and, and predict the outcomes of those March kind of stands. How well we did is still kind of yet to be determined. Some of those teams are still in it. So if we do really well, I'm sure you'll hear more from either Ken or I about, you know, how great the bottle was, if you don't you're too much. It probably didn't do too well. Yeah.
[01:16:52] Ken: Yeah. I think we're in like the top 30 or so percent. And I do have to clarify, Hunter did like 90. 6% of the work I created one future.
So, but you know, I'm happy to be included, you know, in future contests, I will definitely be carrying more of my weight, but yeah, I will. I'm going to share that medium article as well. I actually didn't know you wrote that up, so I'm very excited to share that and yeah, this has been incredible.
Thank you so much, Hunter, for coming on. I, you know, we're going to be talking a lot more. We'll both be at ODS East. I'll probably make a trip to Austin. We can hang out as well. So thank you again for coming on the show and I'm just happy you could tell the story. I think people really enjoy it.
[01:17:36] Hunter: Sure, yeah. It was fun being on here and always a fun time talking to you.