From Navy Electrician’s Mate to Director of AI & ML

Not many forces on Earth are comparable in power to a resilient spirit combined with a strong work ethic, and Dr. Glen Ferguson embodies this dynamic! Tune-in to this episode to hear how Dr. Ferguson went from a young man whose high school guidance counselor discouraged him from even applying to college, to enlisting in the U.S. Navy, to completing a PhD in physical chemistry, to becoming a data science individual contributor, to his present day role as Director of AI & ML at Huckleberry Labs! It is a remarkable story of resilience in the face of adversity and triumph against daunting odds!

FEATURED GUESTS:

Name: Glen Ferguson

Email: gallenferguson@gmail.com

LinkedIn: https://www.linkedin.com/in/glenferguson/

SUPPORT THE DATA CANTEEN (LIKE PBS, WE'RE LISTENER SUPPORTED!):

Donate: https://vetsindatascience.com/support-join

EPISODE LINKS:

The Book of Why: The New Science of Cause and Effect (book recommendation): https://www.amazon.com/The-Book-of-Why-audiobook/dp/B07CYJ4G2L

Deep Learning with Python 2nd Ed (book recommendation): https://www.amazon.com/Audible-Deep-Learning-Python-Second/dp/B09RN7QLT3

Deep Learning (book recommendation): https://www.deeplearningbook.org/

Pattern Recognition and Machine Learning (book recommendation): https://www.amazon.com/Pattern-Recognition-Learning-Information-Statistics/dp/0387310738

Introduction to Statistical Learning 2nd Ed (book recommendation): https://hastie.su.domains/ISLR2/ISLRv2_website.pdf

PODCAST INFO:

Host: Ted Hallum

Website: https://vetsindatascience.com/thedatacanteen

Apple Podcasts: https://podcasts.apple.com/us/podcast/the-data-canteen/id1551751086

YouTube: https://www.youtube.com/channel/UCaNx9aLFRy1h9P22hd8ZPyw

Stitcher: https://www.stitcher.com/show/the-data-canteen

CONTACT THE DATA CANTEEN:

Voicemail: https://www.speakpipe.com/datacanteen

VETERANS IN DATA SCIENCE AND MACHINE LEARNING:

Website: https://vetsindatascience.com/

Join the Community: https://vetsindatascience.com/support-join

Mentorship Program: https://vetsindatascience.com/mentorship

OUTLINE:

00:00:07​ - Introduction

00:01:22 - Glen's first touch point with VDSML

00:03:15 - Glen's military background, college, & grad school

00:12:10 - Glen's first civilian career as a physical chemist working in materials science

00:15:31 - The shift to pursue data science

00:17:37 - A sabbatical in winemaking

00:20:25 - Glen's experience attending the NYC Data Science Academy

00:28:18 - Glen publishes about one of his data science projects in a peer reviewed journal

00:29:15 - Glen's first roles in data science as an individual contributor

00:38:37 - Glen moves into data science management

00:44:12 - Glen's overview of roles in the datasphere and military occupational specialties that map well to those roles

00:55:58 - Glen's current learning focus

00:57:29 - Glen's favorite learning resources

00:59:01 - The best way to contact Glen

01:00:03 - Farewells

Transcript

DISCLAIMER: This is a direct, machine-generated transcript of the podcast audio and may not be grammatically correct.

[00:00:07] Ted Hallum: Welcome to another episode of Data Canteen, a podcast focused on the care of data scientists and machine learning engineers who share in the combine of US military service. I'm your host, Ted Hall. Today I'm chatting with former Naval electricians mate, Glen Ferguson, who is both the V SML member and the director of AI and ML at Huckleberry.

 In this conversation, we talk about how Glen's military experience equipped him with the work ethic, resolve, and resilience needed to navigate discouragement and overcome adversity.

Next, we chat about how Glen discovered a passion for chemistry after exiting the military, how that led to a PhD in physical chemistry and the natural chain of events that took him from physical chemistry. Data science. Next, we discuss how today there are many specializations in the data space. Briefly touching on what those roles require and how some military backgrounds could be great preparation for career in data.

That should be enough to wet your appetite. Let's go,

Right. I'm super excited to have Glen Ferguson on the show today. If you're a member of VS Mail, chances are, you know, Glen, you may know him very well. And I would say the odds are even in your favor that Glen being the extremely helpful person that he is, he's probably even reached out to you and helped you at some point if you're a fellow VDS male member.

And with that, when Glen and I were getting ready for this episode I was asking him, I was like, you know, I don't know if I've ever heard this story. I said, Glen, what is, what was the first, your first touchpoints with the community? And he relayed a very cool story. So Glen, I think that would be a great place to start in this conversation.

If you could just tell us about how you first encounter vds. Sure.

[00:01:43] Glen Ferguson: Thank you. And again, thanks for, thanks for asking me to on the podcast. So the first, first was I was on LinkedIn and I was looking at my network. When you're, when you're looking for people who might be a good fit, your network is always a good place to start.

And this is not something that comes from my network. This was just something that appeared to me in my LinkedIn feed and it was someone looking for a position through V dml and I didn't know what that was. So I, I hooked up v DML from it and I later became a member. But that particular person, I thought they'd be a really good fit for the position.

So, and, and I, I sent their name along. They ended up interviewing. They did very well. And we ended up hiring that person. And then that person has since moved on to a new, a new area of the data sphere. But they did get their start in it through, through vdsl, just reaching out randomly at LinkedIn and finding the right match.

So it was a very advantageous through that I then found the community. I then joined and gradually up my. Interaction with the community actually pretty quickly. Well, I, I thought that

[00:02:47] Ted Hallum: was amazing just because out of the gate, just, you know, when you were looking for somebody that fit a certain profile, you found a VDSL member and like your kind of, your first action as a member yourself was to help another VDSL member.

And then you've, you sort of had a long track record of being that sort of member ever since then. So today you are the director of AI at a lab called Huckleberry Labs, which is super cool.

and, and if you go out to Glen's LinkedIn profile, you're just gonna see accomplishment after accomplishment. Tremendous opportunities that he's had along the way, and he is done very well, from what I can tell in all of them.

But you know, there's a whole backstory, right? You weren't always a physical chemist, you weren't always the director of AI for a an r and d lab, so I'd love to hear that story. Obviously, you're a Vds Mill member. Obviously that story starts with you being in the. I happen to know it started and you know, we're kindred spirits here.

I happen to know, it started with you being an en an enlisted person. You enlisted in the Navy. So I would love to hear the story. Actually even if you could even start before your military time, I'd love to hear, you know, what you enjoyed in grade school. The sort of things you studied where you great at math and science.

Cuz some people think that, you know, you have to love those things or, or have an innate ability in that area to even think about coming into data science. Or did you maybe struggle a little bit more in those areas? And then, you know, how, how that continued to play out as far as you going into the military and all your other pursuits after that.

[00:04:22] Glen Ferguson: Sure. So yeah, so I started out in Indiana, which is very interesting cause I joined the Navy. And there are no oceans near Indiana. Think Michigan does border. That is pretty close, but I'm from as far away in Indiana as you can get from like Michigan. So all the way at the bottom. But no, when I was younger, I, I really did enjoy science.

I really did enjoy math. I didn't necessarily excel in those areas. And I did this all the way through high school. Now I, I talked to a lot of people who take, you know, calculus in high school. That wasn't even an option for us. So when we were in that, when I was in that, it came time that I was starting to graduate and I went to the, the counselor, I think it was my junior counselor, and they said, yeah, you're, you're probably not good enough to go to college.

You should really find something else to do. And I buy a long shot. Wasn't the only person they told that to. That was relatively common. And I talked to other people from other parts of Indiana, they said, yeah, that was extremely common where I was. So so at that point I decided, well, I need to do something else.

I, I knew nothing about college. My parents knew nothing about college. We didn't know where to go. My high school grades weren't exceptional by any stretch of the imagination, so maybe they were going off those, I don't really know. So then I needed to find something else to do. So I looked around and I thought, well, there's no oceans here.

And if, if I join the Navy, I'll get to go see new places. So I did. I also talked to the Marines. I also talked to the Army. Did not talk to the Air Force though. But I, I talked to all of those recruiters and I decided, yeah, the Navy's definitely, cause if I joined the Army, there's a pretty big base not very far from here in Kentucky.

So I don't, I don't wanna come back the ocean, definitely. I'm gonna be far away. So I did, I joined the Navy. Decided to become an electrician. So and then I went from Indiana to Orlando. How long ago it was, we still had bootcamp in Orlando. They no longer have that. I was the second to last company to go through.

And then we moved from Orlando to Chicago where I went to school. Didn't see the ocean, but it was much bigger lake. And then from there I thought, wow, if I do really well in school, I get to choose my orders. But the guy was sick that day, so we just assigned them. So I didn't even get to choose them unfortunately.

And I thought if I get to choose them, I'm gonna try to choose Europe cause I really wanna go see Europe. I didn't get to choose them, so I went to San Francisco. So and I spent the better part of five years on the USS Carl Vincent CVN 70. It's an aircraft carrier as big a ship as you can possibly get.

It's a really interesting environment because. These are very large and require entire huge numbers of people to operate. And if you, if you think you've got pilots, like I think the last podcast was with the pilot. So, and, but the reality is there's tons of people that enable that operation to happen from the checkers who decide if the plane can take off to the air boss who's actually managing the process to the reactor operators who power the catapults, the electricians who manage the 400 vol power 440 volt power to the planes, which was part of what I did.

So, so there's a lot of people that, that allow that system to happen. And I was an electrician, so I worked a lot. I never worked in the 400 vol power ship, although we did work on that a bit. But I worked in all the other shops in the entire ship so large that this is probably bigger. Our division was 110 people and I think that's quite a bit bigger than electricians used that there are like six or seven people.

So so it was pretty large and it wasn't just electricians, it was lots of phone electricians. And of course it's the military, so you have a huge number of jobs. So the guy who slept above me was actually a rescue diver in addition to being a phone electrician. So . But it was a really interesting environment to operate in because I got to, got to see everything as an electrician, you get to go almost everywhere on the ship besides the reactor compartment because everything has electrical equipment everywhere on the ship.

So and there are nuclear electricians. Anyway, we have some of them in vdl, but I was non-clear electrician. A lot of 'em are my friends though. But but that was not where I worked, so I mean, I did have to go and do stuff on the plant because we have equipment in the plant, everyone does. after that, I did get to see some places.

I went on West Pacific Cruises twice to, I spent a lot of time in the Persian Gulf. This was in 1994 to 1999. These were, this was effectively in between the two, what you think of as the golf wars, so after that I thought, well, the officers have a much better life than I did. Now if you look at it now, they basically lived in what would be a really bad college dorm room. But I had 500 people in my space, so it wasn't exactly much better because they have the stacked beds and they're like a maze and there's tons and tons of people.

I guess maybe it wasn't that many, but seems like it, cuz on the other side of our compartment is another compartment with another large number of people. So so I thought, well if I do that, maybe it'll get better. Maybe I can improve, maybe I can do these things. Maybe it's not so bad. So I went back to college and my real, my military experience really paid off in college exceptionally.

Cause when I went to college, I felt like I was very driven compared to everyone else. I didn't feel like I was very driven. I felt like I was just operating at how one should in that environment. But the reality was, I, it was very driven compared to all the other students. So I did, well, I decided to go a different college, so I changed changed my major story that you probably won't hear very often.

I took organic chemistry and I loved it so much. I changed my major to chemistry. So mostly what you hear is the opposite of that. Cause organic chemistry is supposed to be a washout class, but but it, I did, I did well on it. I really enjoyed it. So I became a chemist and later I got into physical chemistry.

So I changed my, I. I finished that and they said, well, you should try to go to grad school. They being the professors, they said, should try to go to grad school for this if you really enjoy it. So I applied, I got into Indiana University. There was a person that I really wanted to work with there. He had just started, he finished at Bell Labs, which was in New Jersey, but it was sort of in the process of shutting down and now it's completely shut down.

But it was a very storied place where they invented things like the transistor. So, and the idea of like the c programming language and operating systems. So I really wanted to work with him. He retired, went to Indiana as a professor. Not very soon after that. Became a distinguished professor, but I turned out to be a second student to graduate.

But I was a seventh student to join US lap. And I think we're seeing again where, because in grad school, everyone would join, was very intelligent, but only two people graduated out of seven. And the reason was we were the people who were the most driven because we were the ones who, we would try something and it would fail, and then we'd just try again.

And that would fail, and we try again until we find a line of research that worked. A lot of the other people would try something sometimes extremely technical, and it just wouldn't work, and they wouldn't know what to do, and they would noodle around for a while and they would leave. So so really that ability to come back in the face of challenges was super important for me to graduate after I graduated.

Quick. Oh, go

[00:11:27] Ted Hallum: ahead. Just to recap, so you, you were originally in high school there in Indiana where a, a counselor advised you to not even pursue college, sort of through cold water on your hopes of hopes and dreams of a college education. So you end up going into the navy, the navy endows you with a work ethic and resilience.

Then you end up getting out of the Navy, go to college, find out that you have a tremendous aptitude for chemistry. You apply that work ethic and that resilience that the Navy gave you, and then it. Ends up making you not just a successful chemistry student, but a successful graduate school student as well.

That's awesome.

[00:12:10] Glen Ferguson: Yeah. So, and I, I graduated during the great recession, probably not the best time to graduate. So I dunno if there's a good analogy, but maybe graduating now. Hopefully that's not the case. But so a lot of people were actually going to graduate school cause they couldn't find jobs.

So we had an influx of people. My advisor said, Hey, you're leaving. This is not the best time, but if you can't find anything else that you like, there's a friend of mine you should talk to. And I went and talked to this person. He was at Agon National Lab as a distinguished fellow who's a fantastic person to talk to.

I really, really enjoyed working. He hired me. I worked there for a couple years. That is a great place to, to work. The National Lab systems were mostly, but not all, started from the Manhattan Project and we actually had people from the Manhattan Project who still worked there, so, which some of them are still there actually.

So which is still today? Still today. Yeah. I talked to someone that's incredible. She said that. Yeah, he's still there. And that was fantastic because talking to them, talking to people at the Wikipedia page is pretty cool to begin with. But then having them sit at your lunch table every day for years is amazing.

It's a great experience. Cause you're gonna ask some really interesting questions like, what was Oak Ridge like when it started? And it's like, oh, I remember when they were putting houses down with, you know, back loaders. So cuz that, that was the time. So but yeah, it's, that was really, was really fortunate.

And when I was there I started really getting the machine learning, because I didn't call it data science at the time, but I started reading papers that use machine learning and I started reading about machine learning other places and I thought, wow, this is really great. What we're doing as chemists is we're looking at, you know, systems of partial differential equations to solve these chemical.

Systems. But what we could do is so much more interesting with these other methods that these statistical methods, which we don't use. So I mean, if we wanna solve a bond energy, yes, we could do that. If you wanted to say, what is the minimum energy configuration of this molecule? Yes, we could do that. But it was very hard for us to move and say, okay, using this aggregate data set, what is the best chemical for this system?

So the idea is that that was not something we could do. I really was interested in that because it was a lot of the problems. We had really moved into that fold. So I started getting grants for, I started not getting grants, but writing grants to do that. And there wasn't a lot of interest at. So, and I think it was really just too early because now it's, it's very popular there.

So, so at that point I thought, I think I'm gonna try to go to a more engineering lab where I do get to work on machine learning. And that's what I did. I left Argon and went to the National Renewable Energy Laboratory, which was made in the 1970s during the energy crisis for obvious reasons, and it's still around today.

And I started working there. Now that's more of an energy engineering focused lab. So I had to do my projects that mostly on batteries and catalysis. And then I was able to do machine learning projects on the side. And I worked mostly with a professor at he was a research professor at the Colorado School of Mines, which is re located really close.

And from there I decided, wow, I really want to get into this. There wasn't a lot of interest in, in the lab system at the time. Again, I think I was still just too early and, and I left and decided to become a data scientist full-time. So

[00:15:43] Ted Hallum: how long then were you effectively, I mean you were using machine learning and effectively doing data science, but not under that moniker.

How long, how long, how many years did you spend doing it? Like in doing data science but in an entirely different field?

[00:16:00] Glen Ferguson: So it was something I did as part of my work for not my entire time at the National Energy Laboratory, but probably about a year and a half. So and sometimes more in depth, sometimes less in depth.

And we used it to answer chemical problems, though we did a lot of work, probably doing a lot more work than we had to cuz we didn't know about the tool sets at the time. So but, But the idea behind it is we just worked in this area at that time and unfortunately that project never, that project got canceled.

And I decided, okay, I want, I wanna be a data scientist. So I left and then started working on data science full time after that to take a sabbatical in the middle of it, which we can talk about if you want. So.

[00:16:48] Ted Hallum: Well, I'm curious, so if you were back in the world of chemistry today would, would they call you a data scientist doing chemistry today or would they still refer to you as some other title, like whatever title you held back in the past?

[00:17:04] Glen Ferguson: To my knowledge, everyone's still, like, they still have very traditional titles like physicist or chemist in the national Lab system, regardless of what you do. So when I was at Argon, I was a chemist. When I was at nre, I was an engineer, so, but it had nothing to do with what I did. A lot of the people I work with were chemical engineers.

Some of them were physicists some of them were other types of chemists. But we all did very similar work. Okay. So

[00:17:31] Ted Hallum: well, I'm, I'm excited. Let's dive into your formal transition to data science.

[00:17:37] Glen Ferguson: Cool. So, so yeah, I decided I needed a little break because I've been doing science for quite some time. If you add grad school together with my time in the national labs, it was 11 years.

So and it's a, it's very challenging to, cuz you're, the main currency of success in that is writing papers. So, which it's not actually finding the answer, it's writing the paper and getting accepted. The, the review process can be very challenging. So and giving to meetings. So at that point I decide to take a little break.

So I made wine for a year. Now

[00:18:12] Ted Hallum: that's fascinating. I mean that is not a normal pastime that a lot of people get an opportunity to pursue. So the chance that you got, the fact that you got a chance to pursue wine making for quite some time, that, that's awesome. But I'm curious at that point, were you also applying data science methodology to like the wine making process?

[00:18:33] Glen Ferguson: No, what I mainly did was look for a way in. So what I did when I went into data science was this, or when I went into wine making was the same way as I did to everything else. I tried to find experts in the field and talked to them about it. So, and then when I found them I was like, okay, what's the right way to do this?

And I then followed that way to do it. But I was reading, trying to figure out how you were getting it. Cause I wasn't getting a lot of traction in data science in the national lab system. So I thought how do we get it in business? So I really spent a lot of time trying to figure that out. And also ever since then, I still read about machine learning all the time.

So when I went to, I did a bootcamp after that, and when I did the bootcamp, they, they said, oh, one of the really good books in this is Introduction to Statistical Learning. So, which is a good book or I think it's maybe called Elementary Statistical Learning, but I already read it. So so it wasn't that, that big of a challenge.

So even when I was making wine outside of the harvest season's, very intense. But outside of the sites of that, you have some time. And I was reading about data science, making sure my python skills were up to scratch, which was not a problem.

[00:19:38] Ted Hallum: before we move on and hear about your bootcamp that you went to and everything like that I, I do have to know, are you still into wine making on the side?

Can we get a, a bottle of your vintage or how, you know, what's the status of that?

[00:19:51] Glen Ferguson: I do not. It's, it's something that, it's very challenging to do as an individual because it requires like a lot of time, a lot of effort and a lot of space, which I don't have. I mean, I live right outside of New York City and it's very hard to get into this space.

It did get into home growing afterwards, which I still do sometimes. So but even that requires a relatively large amount of space and an apartment near New York City really isn't the best place for that. So still do have some friends who are into it though, so, and they, they do have vintages, so.

[00:20:25] Ted Hallum: Very cool. Well, hey, if your situation ever changes and you have more room and you get back into it, keep us posted. We'd love to get a bottle. So you mentioned that you went to a data science boot camp to really sort of formally start going after data science in a, in a disciplined way. What did that process look like for you?

Because I'm, I'm guessing that compared to now, that was sort of the early days of data science boot camps. It might have been a little different than what someone might encounter if they did a boot camp today.

[00:20:55] Glen Ferguson: I think it's vastly different. So and at the time there were boot camps that accepted PhDs.

I think the data incubator is the name of it. They did not charge you to take the boot camp, but they ch they got. Their pay out of the job you got afterwards. And I think that model is still around for some of them, but I'm not sure. But I think anyone can apply to these bootcamps now, and you just, if you, you can, it depends on which track you get accepted into the no pay track or you have to pay to go into it.

So at the time though, there was a bootcamp, I didn't know about that bootcamp, and I ended up going to a bootcamp that was just a normal bootcamp. I tried to find one that had good entrance criteria and had people who actually knew about the subject that they taught. So it was run by people who went to Brookhaven, I think Stony SUNY Brookhaven.

And so it's state University of New York, Brookhaven and, and other colleges. So they were reputable and. I talked to them, I got a relatively good feeling about it. The truth about boot camps though, is even if they have instruction, the real meat of the boot camp is doing projects on your own. So if you said you could replace a boot camp with just meeting 20 or 20 or 30 new people who are all really passionate, wanting to go into the subject and giving you a lot of projects that you need to work on and help whenever you need it, that's probably as good as it gets.

So that's really what it is, is people are really, you know, I got more out of the student interaction than I did out of a lot of other parts of it. So I'm not, they did have didactic construction and some of it was good. Some of it was very, very high level. Other parts of it were more challenging.

But so in that point, from that perspective, boot camps, and I think this is how the data incubator was set up, where you just got projects and you had to produce output at the end. There was no form. Some of them had no formal instruction. I don't know if that's true of any of them anymore. Also, the field has grown immensely in the time period since then.

And then everyone was a data scientist, like everyone. Now very few people are just, just data scientists. They have data m, machine learning engineers, ML ops engineers, data engineers. None of that was, was part of the equation when I'd, data engineers are probably the first, but you know, lops I think came around last year.

so it's, and the number of boot camps has also grown a lot. And there's also graduate programs and I don't, I think the graduate programs might have just been getting started, but I don't remember them being super significant at the time. Now the

[00:23:30] Ted Hallum: bootcamp that you did was, was that the New York City Data Science?

[00:23:35] Glen Ferguson: It's, yeah, the NYC Data Science Academy. Okay. So, and they are still around, so,

[00:23:40] Ted Hallum: and, and you did it, what was that, 2016?

[00:23:45] Glen Ferguson: I believe so. I can't remember. So I started, might have been started in early 2017, so, okay.

[00:23:52] Ted Hallum: Just so people have a point of reference in time for, you know, when we're talking about when they're trying to maybe size up Yeah.

Their current experience or their experience two years ago to, you know, to the experience that you had. Yeah.

[00:24:05] Glen Ferguson: And yeah, it's, it doesn't seem like it was that long ago to me, but it, I think it has grown immensely since then. or the, as a profession, not the MIT Data Science Academy, but the profession as a whole has grown so much in the last few years.

[00:24:20] Ted Hallum: So, and so given the state of things at the time I take it that you probably weren't doing a lot of Setting up APIs to deploy models or anything like that is probably more core data science. So how do you feel like it, it did in terms of preparing you for that capacity and then how well did that translate into actually getting a job

[00:24:42] Glen Ferguson: afterward?

I think the biggest, because a lot of it has to do with how motivated you are. So if you're talking about me, it probably helped me out a lot cause I was able to really make a good network, talk to a lot of people, do some interesting projects, form that into a portfolio. Not everyone was able to do that.

So so I think it has, a lot of, it has to do with the motivation of the student a lot, but a lot of the bootcamp, and I don't. What the current state of it is. But the help they give you is by either writing your letters of recommendation or connecting you to jobs afterwards. And that could be very significant.

And I know that some of the boot camps, that's a huge part of their business because that's how they get paid. So is they, you can do part of your pay. And I don't think they're alone in that. I think some of the coding boot camps are the same way. Like I think the Grace Hopper Academy might be like that.

So

[00:25:36] Ted Hallum: my takeaway in how you described your boot camp experience is that and, and all of us who have served in the military know this you often in, in life get out of an experience, what you invest into it. And it sounds like that's exactly what you experienced at the

[00:25:49] Glen Ferguson: bootcamp. Yeah, even more so than I think that's true of college too, but it's more so with college because the boot camps, the model is ever changing and it's super complex.

Because the. It's hard for them and they can change their curriculum every couple of months. It's hard for them to keep up. So, because the professional needs to change on a nearly weekly basis sometimes. So a

[00:26:13] Ted Hallum: lot of times it's not about you just doing the assignment and checking the box, it's about going the extra mile and thinking, okay, this is what they're teaching me, but what else do I really need to know?

And going out on your own and learning

[00:26:26] Glen Ferguson: that stuff. Exactly. That's because you have to go out and say, okay, well this is how I'm doing it here, but how is it done in practice? You know, there's a big difference between making something really interactive and shiny and what you're actually gonna do in practice.

And a company, does it even need this? Does it just need an api? So who am I communicating with? How is this being, how is this data gonna be used? How is this data gonna be ingress into my system? What are the challenges with data typing and things like that? So And really understanding that you start to really branch out into software engineering, because as a chemist, I didn't think about data types too much because most of what we deal with is numbers.

So, and, and neural networks, like I'd heard about them before I actually read about them as a chemist, believe it or not. There was a, something called a hot field net that I read about. I remember this was in grad school and that was sort of a rudimentary neural network. I didn't know it at the time, but now I recognize it as such.

So, and a lot of this, the math has been around for a very long time. So for neural networks, it might even be like the forties or the fifties. So, but it's, it's been quite a long time ago.

Yeah. Yeah. And the first time I ever heard of these in practices when I was in grad school, I heard about someone I was talking to, a place that did cold storage and they're like, oh yeah, we have our thermostat operates on a neural network. We don't really know how it works, but it does. So but yeah.

[00:27:54] Ted Hallum: The data scientist who built it would probably say the same thing. Yeah. I tweaked though though. I, I endlessly tweaked the hyper parameters and it works great, but I don't know how Yeah. So I, I take it from a comment you made earlier that the flash to bang from you doing your bootcamp to having your first data scientist role probably wasn't too, you know, that fuse was pretty short.

What, what did that transition look

[00:28:18] Glen Ferguson: like? So I finished the bootcamp and then I continued to do projects and do work. One of the things I did when I was so. For me, I have a pretty high bar for things, so I, to feel like a real data scientist, I wanted to publish something in a journal. I don't know why you'd wanna do that now.

Now I would never say anyone would need to do that, but that's really the way I felt at the time. So when I was in the bootcamp, I started reaching out to people and saying, Hey, can we do some work in this? You know, I have this data science work, I'd like to work with you. And the one person that responded was a Metropolitan Museum of Art in New York City.

So they said, yeah, let's talk. So I went and I talked to them and we did a project, worked out very, very well. And I published that in the Journal. Now I did, I did that after I got in my first job, but I was working on the project the entire time. So and I would guess I finished the bootcamp sometime in, in the in the spring.

And I got my job that. So nice. Fortunately the first job as a grant manager, which was precisely what I was looking for, cause you get these grants, you'd be like, wow, this would be great to do, but I don't get to do it. I just gave, gave a grant for it. So, so I moved on from that cause it was not a good fit for what I wanted to do.

And then I moved into being into a hospital. So

[00:29:36] Ted Hallum: now I'm glad you mentioned that because I saw those, I saw two, at least two roles that were sort of medical related combine or, and, and data science. So I was curious if that was just sort of a happenstance turn of events or is, do you have something in your background that specifically kind of took you into those medical, those roles that combine medicine and data

[00:29:59] Glen Ferguson: science?

I don't know if there's anything specific in my background. I could kinda see it because I started out as a health professions major in college. And I did do a little bit when I was in the Navy. I used to go talk to medical all the time. And then my father-in-law's a doctor now. So, so I do see a little bit in my life that does sort of focus me towards that, but there's, I mean, there's nothing, and I'm, one of my degrees is in biology, so from undergrad, so, so maybe I had a little bit of a setup for it, but, but no, I didn't have any medical experience previous to this.

So it is interesting. It's nice because one, it's an area of enormous need, which is important. I mean, data science can really offer a lot in medicine. And it's an area that has lots of data, so you have good data, huge need for it. So I, I think it's an area that's really ready for data science as opposed to a lot of other areas, which I'm not, I'm not sure exactly are at this point.

[00:31:02] Ted Hallum: So Well, I can definitely see where some of those things you mentioned were probably an initial little burst of energy that kind of pushed your career in that direction. But then I'm curious, once you had the first role that sort of combined data science and the medical world, did you find that that sort of snowballed into the next roles that you had?

Where it's like, okay, you've, you've done this combination of of the medical world and data science and then that kind of catapulted you like you were picking up momentum almost in that direction.

[00:31:37] Glen Ferguson: Yeah. And if things had turned out a little differently, I'd still be at that first job or that.

I think it was actually my second job official. I'd still be at that second job. So, cause I really enjoyed it. It's, it's really fun. It's great. I love working with doctors, they're great people to work with, so and I knew I was teaching doctors and that's, which is really interesting. So cuz they're like the best students you'll ever have.

because, you know, some of times they're coming back, they work the night shift and they're coming back from that and they're still super engaged in their class, so it's, it was really fun. I really enjoyed that. And actually there's probably more, but I'm still under NDAs from some of, so I can't talk about it, but Sure, sure.

But yeah, I, I really enjoy working in, in different aspects that are appeal to human health. So because I do think it's ready for it, it's so ready. It's something where you can make a huge impact in data science. And this is, There's a lot. I mean, right now a lot of data science goes towards advertising, so, but I really feel like, you know, if we could really help people out.

Cause if it's the health, your health is really important and data science could really offer some great help there, maybe not. And I, a lot of people talk about like automated diagnosis. I'm not sure that's where it's gonna offer the greatest help. You know, a lot of it, the reality is that say like, let's, let's talk about radiologists specifically.

If you're a radiologist, you spend most of your time looking at normal tissue, which maybe that's interesting at first, but it's really not helping anybody, you know, to say we don't see anything in this x-ray. But when there is something to see, you want them to see it. And if what data science could do, what we felt was if it could accelerate that time, and this is one we never realized, unfortunately, if it could accelerate that time to finding what they need to and giving radiologists that.

That's a huge benefit as opposed to recognizing something in an x-ray, which they can do perfectly well. So, and they can do many thousands more than in a model that does one. So well, Glen, I

[00:33:36] Ted Hallum: know that you're all about helping people with their careers, particularly helping fellow veterans, and that's something we'll talk more about in a little bit later in this episode.

But I think there are two big takeaways to help people with their careers that, that you actually just covered. And the first one is that the medical industry is such a rich place to do data science because they have, as you mentioned, data, there are lots of fields where not so much data is available, doing good data sciences, more difficult.

And then you mentioned the fulfillment factor too, right? There's only so much personal satisfaction that can come from helping companies. Products that people don't need in most cases. Right, exactly. Whereas, you know, you're making a very real, tangible positive impact if you're taking data science and, and bringing it to Bayer in the medical world.

But the, and the other thing that I think you mentioned that's key to careers is people should be mindful about where they're making investments. Not just their technical skills, but in their domain knowledge as well. Because once you have a role, particularly if you excel in that role you do start to pick up almost like a, it's like a momentum or a gravitational pull.

And so if it's something where you're interested and you're passionate, then yeah, make those investments. Maximize that momentum. Keep your career going in that direction. But if you're getting into it and you're finding that it's really not something that you love, then I would caution. Investing too deeply.

Try, you know try to fail fast as they say, and move on to a different opportunity. Not actually fail, but you know what I mean. Yeah. Decide that it's time to cut sling load and move on to the next thing before you invest deeply. Because the deeper you invest that, you know, if you're there in that role doing good things for years, what you're gonna find is that in the future you're like a beacon for recruiters in that world and everybody wants to give you a job continuing to do the thing that you don't love.

So yeah, just be prudent in how you make those investments. And if you're not loving it sooner rather than later, move on to something that you do love. Right.

[00:35:45] Glen Ferguson: I think that's really good advice cuz it's very, I mean, if you do it does, it does tend to snowball a little bit. So because people will pick up that you're interested in this area and that maybe there's, the way hiring is done now is with a lot of it your behavioral hiring, which really favors people who've done the job before.

So makes it a little challenging. I think that's especially challenging for veterans because if you're, if you're in the military, you're doing things that are challenging with data, but a lot of times it's very hard for people to see the connection between intelligence data and customer data. And I don't, I don't know why, but which is probably why I picked up.

On the vd vd DSL member that was posted on LinkedIn. Cause I looked and I thought, wow, this is a fantastic background for someone looking to get into data. They have a background in human intelligence. And I thought, wow, this is someone we should talk to. But that was, that's, you know, they've been looking for a while and they had not found that.

so I, think that that communication, trying to communicate your interest area and your skill set into, or from the military into the civilian world's, very, very.

[00:36:53] Ted Hallum: Yeah, absolutely. And I can't wait to dive into that more in the, like, the nitty gritty with you here in just a little bit. Cause that's gonna be interesting.

But before we do that you've told us about your first data science roles after the bootcamp. Those were more, sounds like more of like individual contributor type roles. Obviously your role now is more of a leadership, you know, you're managing people who, who are doing data science and machine learning.

So take us down that path real quick. What, what did that transition look like going from individual contributor to managing individual contributors?

[00:37:29] Glen Ferguson: So , I, I don't wanna promote that this is because a lot of people, like in some professions, this is the only way you can go. You, you start out as an contributor and then you become a manager.

As a data scientist, like a software engineer. Maybe that's not the case. Maybe you can stay individual contributor and still make really strong contributions. Now, What happened with me is I be, I was a consultant and I was a data scientist and I thought, really, I've done very well at this. I, I wanna move up to the next level, next run of the career.

They were not interested in that, so so I left there and moved on to another, a job at a software as a service company. And there I moved into a senior data scientist role and then later into a lead data science role where I was, I was helping a team and I thought, wow, this, this, I feel like this is going well.

But there are some business decisions that I did not think were the wisest and I moved on from there to a consulting role again. So in that second consulting role as a principal data scientist and I, you get to see a lot of companies from the inside there. And what I've seen my entire data science career is that we're looking at a disconnect here between the people managing data scientists and data scientists, and I thought you.

I wanna see what it's like if that's not the case. So I then moved, and that's into my current position as the director of data science, which I kind of feel like was a lateral move from a principal data scientist. As a principal data scientist, you're a technical manager and you could move up from there to be like a chief data scientist, where you're really solving sometimes academic or semi-academic problems in data science.

Like how is it possible to do this? And as, but as a director of data science, you have to spend a lot of time working with people and trying to figure out how they can, you know, gear their careers and sometimes challenges they have and how can you support and help them really achieve what needs to be achieved.

And then that's with the, the group members directly. And then outside of that, it would be in marketing, the group helping develop the roadmap, selling the roadmap to management. So so there's a lot of. Soft skills involved in that process. So that's, so that's sort of how it worked for me. I don't know how it works for, for others in that position.

I could see, because I mean, the idea that you have, if you have 23 years of leadership or 23 years in the military, and a lot of that's in leadership, going to an individual contributor role for a while doesn't seem, might not seem so attractive, but in data science you really need those technical skills because as a director, I still write code.

I was writing code earlier today and I'll write code later today. So I think it's really, I think data science is one of those areas where hands on keyboard doesn't go away for a very long time. So my new people are VP used to still write code. So and part of the deal is I don't write as much code as I used to by a long shot, but but the idea is that it never really goes away.

But a lot of the code I write right now is to help other people out. So like, if they're writing some, a big program I'll try to take advantage of my, my coding skills to write things that make it easier for them to do so are easier for them to write the bigger, the bigger system and I'll, I'll say, okay, let's, I'll write this, this lower level system here.

So you, you don't have to worry about it. So so, cause I'm, I'm, I'm trying to help empower those, those people who are working in that area. So, and sometimes it's, you know, sometimes it requires a higher level skill set. I have to take advantage of that as well. So, but yeah, that's, it's definitely a managerial role and I think there probably are some data science directors who don't, so, but I, I kind of see it like data science is, is, it's new enough that you still get hands on keyboarding experience while you're a manager.

So if you have a lot of leadership experience and you have really good technical chops, it's probably something that people could, that the profession could really benefit from.

[00:41:17] Ted Hallum: Yeah, absolutely. And you mentioned how such a big component of that role is supporting and enabling the people who are subordinate it to you and.

If your, if your people underneath you are data science individual contributors, it's gonna be really hard to, for you to effectively support and enable them without helping out with some code sometimes, or at least helping with code reviews and that sort of thing.

[00:41:39] Glen Ferguson: Yeah. Yeah. And it's I, I feel like that's a, a very strong component of it and that's, I feel like that was definitely something I learned when I was in the military.

One of the things that I learned was, you know, people need support, you know, because sometimes they might just be kind of afraid to do something. Like people are afraid to go into the, the plant. Sometimes they're, they're afraid to go on the mass, so you have to go with them because it is legitimately frightening.

So, you know or even on an interesting area with this, I never thought of, you know, how do you do how sometimes you have to fix stuff on the flight deck during flight operations? So they have to suspend them while you're doing it, which means it has to be done really quick. It has to be done right. It has to be done fast.

So and there's no. manual or procedure for having that happen. A lot of times you have to do it in the moment as you do it. So how do you support people so they can believe in themselves to do that? So so with that, I think that really goes through all the way till now when I'm working with people to deploy a new system.

What's the right way to do that and what's the right way to support that?

[00:42:44] Ted Hallum: Absolutely. Well, before we change gears and move to the next portion of the episode, I just wanted. A quick pause and thank you for being so transparent about your background, about some of the discouragement that you encountered initially, you know, as early as high school the different adversities you faced, and then sort of the way you maneuvered that to get to where you are.

I I hope that that's a story that our listeners can be encouraged by. I know that a lot of your background is gonna resonate with many of the people in, in our community, and so if they're sort of at a point in their path where they're stuck, then I hope that they can take something you said and get unstuck.

And I can, I can definitely see where that might be the case. So now thinking about since you joined the veterans and data science machine learning community you've found ways to contribute all over the place. But there were some key ones that I thought were really interesting that I wanted to talk to you about.

I'll start with the most recent thing. Very recently, you. Wrote an article that we published on our website, and then you've, you've authored quite a few other posts that have, have been in our discord. So if you're not in our discord just know that you're missing out on some amazing thought leadership in this area from people like Glen who are putting things out there that will really help you across the full breadth of your data journey.

But I wanted to to talk specifically about a post you had done entitled, transitioning to Data Professions from the Military. Cause I thought it was super interesting you went through each one. So there was like data engineer, data analyst all the way through like lops engineer, and you took each of them and you broke it apart and you said, okay, well what's required to do this role?

And then what sorts of backgrounds do people have coming outta the military? And then you kind of did like a little mapping of. All right. You know what, what sort of prior service members fit well as data science researchers or machine learning engineers? That I think that article is just fascinating.

I think it's tremendously helpful for so many people in our community. And it sort of ties back into some things you've already hinted at earlier in the episode. So I'd like to just go there, if you can kind of take us through that article and, and those different mappings that you had laid out.

[00:45:08] Glen Ferguson: Yeah. This is, and this is really interesting because I've been trying to, I've been thinking a lot about this over the last couple of years about where everyone fits in because calling everybody a data scientist, this doesn't seem appropriate anymore. Cause people's jobs are so different. You know, and this will vary because I talked to some people who just.

Scala for using Spark to pull very, very large data sets. That doesn't seem very data science to me, that's maybe more of a data engineer or maybe a very specific one at that point. So the, I, so for me it's very sort of breaking them down and I sort of, the way we think of data scientists, I sort of think that's more like a data science researcher because that's someone who's going to take like this way of doing it, which is gonna be, okay, we have a problem.

What is the right way to solve it? There are some, there's some information here probably, and this is something that is really good for scientists or engineers because they're going to approach it more like a scientific, and I say if you're, if you're into advanced engineering, it's not that different from science.

They're going to approach it like that as a problem in and of itself in the process of doing it. So they're going to say, okay, this is the right way to approach this, this problem, starting from the data. So I really do think it's probably easier to do this if you start from a PhD level in science or engineering, which is probably why I meet so many people from those areas in data science.

And so many people are from those don't find any trouble with it. So if you're someone who has that sort of mindset or you've been doing it before, I think it's very easy. I think this is really good for people, our intelligence to go into because they're used to approaching problem or data with a problem mindset.

You know, what information do we have or what do we not have and what do we need to know? Maybe also from electronics engineering. So because it's, to me, it seems like they don't seem like they have a lot of skills overlap, but you seem to meet a lot of people who have skills overlap in these areas. So if you really want to get a PhD or.

I wouldn't discourage it, but I don't think it's required by any stretch. But I do see a lot of people who use, and this is probably good and bad, I see a lot of people from chemistry, especially theoretical or computational chemistry, moving into data science. And I think the reason I see that is because there's so much overlap to them to begin with because you're used to using math to solve with data, to solve problems.

And the problem with that is now we're losing people from that profession, which they weren't necessary to begin with because a few years ago you would've thought, well, maybe what's it used for? But drug design, I don't, I don't know any drugs or design not using computer now. So so you have these drug design companies, they're a lot of people on them used to be or are computational chemists.

I could refer you to some in New York if you really wanted, but, but the idea from it is that area is what we think of traditionally as a data scientist. If you think of data science in your head, that's probably what you think of. So and a lot of people from the natural sciences really move into this very easily.

If you're talking about machine learning engineer, which is another really important profession, these are people who are gonna take sort of after that scientific problem or engineering problem has been figured out and they're gonna transfer it, it's a different problem and transfer it into a reusable set of code.

So you can take that code set and write it out into something that's scalable, reusable and effective for what it's trying to do. And this is where you're gonna get into building these APIs. This requires a little bit more of the CS knowledge and maybe it's a little bigger of a transition. And a lot of people probably do have backgrounds in CS who do this.

So if you're talking about parts of the military that do use, like if you're a. This was an option when I joined to actually ask if you wanted to be a cold ball programmer, which at the time that seems like an archaic language, but I don't see who it is because after, after I left the military, I programmed in 4 20 77, which can't imagine much older than that.

So but, but yeah. So if you're some from some, a position like that, that seems to be really, would be a really good fit for it. Or if you're someone who just has been doing this, has been I, I can imagine, although I, I don't have any direct experience with this people, someone who's like a CB where they're used to building things, you know, from something that's been done before.

CB is a, a builder from the Navy. For anyone who's not not aware, so that seems like it would be a pretty good fit to me. Probably the easiest entry point is gonna be a data analyst and originally a data analyst. They sometimes are called business analyst. And I think the difference is that business analysts get paid more.

But they don't really seem to do much different. But that seems to like a good entry point for people because if you have like a good skill set in understanding data, it's something that seems reasonable to go into. It's not something that's gonna require really in depth knowledge. And if you're, if you're an officer and you're doing, you probably do a lot of analysis now cuz you have to make decisions and if you want to use data to make those decisions, that seems like a reasonable overlap to me.

So. The next one is data engineer. Data engineer probably is more difficult to go into cause a lot of data engineers I meet seem to have a lot of background in computer science. So because it's something that's more computer science heavy, cuz you have to deal with SQL and no SQL databases, very high levels perhaps building them.

And it's something that's more, more challenging if you don't start with a background in software engineering. So if you have a military rating, you feel it qualifies, qualifies you to move into software engineering. That's probably a really good start. And data engineers are it is a very good place to enter the profession right now.

The last, and I don't think it's actually the last, but the last for now is machine learning ops engineer. And this is, It's gonna involve a little bit. It's more, I thought it was gonna be just a specialty area of machine learning engineer for a long time, but seems I was wrong in that. Because really what it did is it developed a lot of specialized skills that seemed to be more of DevOps like infrastructures code C I C D, and then those are very much software engineering skills.

But then now it's bled into how do you deal with models in production? How do you deal with automated retraining of those models? How do you deal with making sure those models have appropriate observability? And this is really a newer area. There's not as much established with it, in my opinion. And I dunno if you feel differently about this cause I know this is what you work in.

So but, but those are sort of the areas I think the key is, What areas or what skills do I feel like I have now that have some or good overlap with these areas and how do I think I can put them into, into these areas? Data engineer being sort of the most difficult at the moment. And maybe allows me easier cuz I honestly don't think a lot of companies have a good understanding of what they need out of lops engineers.

So, which is unfortunate, but where we're at right now.

[00:52:05] Ted Hallum: Yeah. So I agree with you on all of that. I feel like if business people do their due diligence to understand at least at. Something more than the surface level, how all these things work. And if they're desiring excellence out of their particularly for machine learning, then they're gonna know that they really, that all of these things are discreet and they need experts in each of these areas.

I agree with you. I think lops engineers already are separating themselves out as a separate discipline from machine learning engineers. They're really that, those infrastructure experts that make sure that everything is designed and spun up so that the machine learning engineers have the tooling they need to do what they do.

But that, that leads me to a question I wanted to get your thoughts on. A dynamic that I feel like I perceive out in industry, I feel like that for a certain portion of, of industry, there are business people who feel like data scientist, machine learning engineer, lops engineer. Nah, all data people

What I really need is a software engineer who can do data science and machine learning good enough to get a 60 to 70% solution into production. And so they really don't seem to have much interest in having deep data science expertise or deep data engineering expertise. They just have moved into this mode where they are, that they know software engineers can do production code.

So they're just looking for software engineers who can tack on mediocre at best model building skills and get stuff out the door. Do you think that perception is right, first of all, and and what do you think the implications of that are? Because I can see there being some really bad ML outcomes for people that are choosing to pursue it that way.

[00:54:01] Glen Ferguson: I think that's a good way to end up with a lot of models to degrade very quickly, if you're lucky. So but it's something I've talked with about software engineers who now work in data engineering and they seem to think, well, there are, there is DevOps in software engineering, but when you move into machine learning, the data engineering that you're doing, it's not the same as backend engineering as if you're a software engineer.

The DevOps that you're doing for, which is now ML ops isn't the same. There are specialized things in it because you're dealing with the data differently than you used to. So if you're dealing with the data differently, because now data's different. It's not just before. If you're just looking at like your amount of web traffic, that's, that's different that you can just look at the web traffic.

But if you're using the amount of web traffic to predict something, you have to make sure that that web, the data coming into it doesn't have a lot of gaps if they're important, doesn't have a lot of missing data or incorrect data. So those specific data operations. They're probably not gonna get done very well.

And this is why I'm not actually, boost is one of the things that annoys me because it's used all over the place because it works really well in Kaggle and it works beautifully. So, and I don't have anything wrong. I think it's a fantastic model, but I feel like it's a go-to for a lot of people. And it's not because they understand what's going on, they just feel like, oh, let's use XD Boost.

So there's a, an article I read on Medium was really good. I think the author Susan Lee, but she, it was on an article on linear regression versus neural networks. And she does for, I think it's just a feed forward neural network. And she does both. And she shows, well, you get kind of the same answer, why don't you just use the simpler model that's more interpretable?

And I think that's really important. So because that's something you get from data science, it's not a software engineer, just using an ML ops tool isn't gonna get.

[00:55:46] Ted Hallum: Right. No, I think you could not be more correct. So Glen, I think at some point in the future we're gonna have to have you on for another episode because there's so many things I wanted to talk to you about today and just needed time constraints.

We're not gonna be able to talk about all of them, so I hope you'd be open to that. But before we let you go, I definitely wanna ask you some wrap up questions. Things that people typically want to know about our guests that come on the show. So obviously you are the product of continuous lifelong learning.

You've been learning all along the way, and I'm sure that you haven't stopped anytime recently. So what's your current learning focus? What are you curious about and poking around with now?

[00:56:25] Glen Ferguson: So the areas that I'm really interested in now, cause I, one, I think they're gonna be really important. And two, they're just fun to read about.

Reinforcement learning, I think is a fantastic area that has not really read to its full potential. Everything you read about it is about how good it does in video games or playing tag, and it's like those, those are not that interesting in reality. So once it starts to reach actual business focus, it becomes really interesting.

And I'm trying to figure out where that is beyond contextual bandits, where it is, it has already shown it. Another area is causal inference. Cause one of the biggest things we're missing in machine learning is we don't know why things happen. So we can say this happens from these variables, but we don't know what variables drive that to happen.

And I'm a huge fan of the book of Y by G Pearl. I'd rec highly recommend that book to any of any of the listeners. But really sort of understanding why these things happen I think is a really important part of machine learning that we currently aren't using till it's, it's full advantage.

Those are two, I have a lot more, but those are. We'll stop it here.

[00:57:29] Ted Hallum: Hey, that, that gives some great starting points for people who maybe have just finished learning something and they're like, well, what's the next thing I'm gonna dive into? There you go. Go after these two topics that you've gotten from Glen next Glen, I was curious about as far as resources like books what, what has been say in the last six months on your radar that you'd like to recommend to the audience?

[00:57:54] Glen Ferguson: Yeah, this is, so first we'll talk about books. I think it just came out in its second edition. It's called Deep Learning with Python by Fran Fa. But it's not, it says Deep Learning in Python, which makes you think it's gonna be all about neural networks. I think it's a fantastic introduction to to machine learning as an idea.

So I really enjoy that one. There's a book called Deep Learning by Ian Goodfellow, which you can read for free online if you just type in like, deep learning book. It's from MIT Press and they, they don't give you a pdf, but they let you read the entire book for free. So I would, I would highly recommend that one as well.

Or if you wanna pay for it, you can. Then this is an old book, but I've, I really like, and everyone I've ever talked to really like it, I think it's called Pattern Recognition and Machine Learning by Christopher Bishop. So, and it's a really good book on understanding, especially all of these methods in machine learning.

It's really good at it, and it's the second best book I've ever read. Close to Introduction to Statistical Learning, so, which is a classic.

[00:58:52] Ted Hallum: So I've heard good things about all of these books. The first one, deep Learning with Python. I would add that the author is also the creator of the Carus package.

Yeah. The Carus api. So if you've used Carus and you like it, well, you should know that this, this book recommendation is actually from the guy who created that. So I think that's pretty cool. Well, as we wrap up here, Glen, I've had your LinkedIn username on the screen throughout the whole episode. Is that the way you prefer to be contacted or is there any other ways that you like to for people to reach out?

[00:59:23] Glen Ferguson: That is a great way to contact me. You could also email me and it's just gonna be my full name. It's like what you see on the screen. Plus it's Glen dot Allen Alan, a l l e n ferguson gmail.com. So those are the easiest ways to contact.

[00:59:40] Ted Hallum: Outstanding. Well, Glen, thank you so much for your time coming on the show.

Thank you for sharing with us your story and your background and these insights about the various roles in the data space. And then, you know, your thoughts on resources and things that you wanna learn. I think it's in its entirety, it's a source of inspiration for members of our community. And I've just enjoyed talking with you.

So thanks for coming on and I do hope that you'll come on for another episode in the future because I've already got a bunch of notes of things I wanna talk to you about.

[01:00:12] Glen Ferguson: No problem. Thanks for having me on. I really appreciate it.

[01:00:14] Ted Hallum: Thank you for joining on this fantastic conversation with Glenn. Until the next episode, I bid you clean data, low P values, and God speed on your data journey.