The Impact of VDSML’s 2021 Scholarship

Sam Sipe was a naval aviator, VDSML's 2021 FourthBrain Scholarship recipient, and he's now a budding ML Engineer! In this episode, Sam and I chat about his childhood interest in rockets and flight, which led him to pursue an education in aerospace engineering and military experience as a pilot. Sam tells us about how his engineering pursuits initially sparked an interest in computer programming, which eventually grew into an interest machine learning. Soon after exiting military service, Sam won VDSML's inaugural FourthBrain Scholarship. Sam tells us about his experience with FourthBrain's bootcamp for aspiring ML engineers - as well as some preliminary information about VDSML's 2022 scholarship opportunities!

 
FEATURED GUESTS:

Name: Sam Sipe

Email: sam@sipe.io

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

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

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

EPISODE LINKS:

Sam's Personal Website: https://samsipe.com/

Sam's FourthBrain Capstone Project: https://amplifygrid.com/

Sam's GitHub: https://github.com/samsipe

Lex Fridman Podcast: https://lexfridman.com/podcast/

Thinking, Fast and Slow (book recommendation): https://www.amazon.com/Thinking-Fast-Slow-Daniel-Kahneman/dp/0374533555

The Signal and the Noise (book recommendation): https://www.amazon.com/Signal-Noise-Many-Predictions-Fail-But/dp/159420411X/ref=tmm_hrd_swatch_0

3Blue1Brown (YouTube recommendation): https://www.youtube.com/c/3blue1brown/featured

pyimagesearch (Learning Resource Recommendation): https://pyimagesearch.com/

FourthBrain (Bootcamp Recommendation): https://www.fourthbrain.ai/

DataCamp (MOOC Recommendation): https://www.datacamp.com/

 
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

00:02:17 - Sam's military background and personal data science journey

00:17:02 - Sam's thoughts on the value of a background in engineering

00:19:06 - Sam's description of the FourthBrain bootcamp experience

00:23:58 - Sam's goal for doing the FourthBrain machine learning engineer bootcamp

00:25:57 - Sam's epic FourthBrain capstone project

00:35:49 - Challenges Sam encountered during his FourthBrain capstone

00:40:05 - Preliminary info about VDSML's 2022 scholarship opportunities

00:47:41 - What does Sam plan to do with all these newfound skills?

00:54:07 - Sam's current learning focus

00:56:16 - Sam's thoughts on cloud service providers

00:58:42 - Sam's favorite learning resources

01:03:26 - The best way to contact Sam

01:04:40 - 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 the data. Canteen, a podcast focused on the care and feeding of data scientists and machine learning engineers who share in the common bond of us military service. I'm your host Ted hall today. I'm chatting with former Naval aviator Sam site. Who's both a VDSL member and a space systems engineer, a technologies.

In this conversation, we talk about how Sam's childhood interest in rockets and flight led him to pursue in education in aerospace engineering and military experience as a. We also discussed how his engineering pursuits initially led to an interest in computer programming, which then grew into an interest in machine learning.

And that leads us into the topic of how Sam came to win VDS. L's 2021 scholarship to fourth brains bootcamp for machine learning engineers. You'll hear about Sam's experience with that as well as some preliminary information about VDS L's upcoming 2022 scholarship opportunities. I think that's enough information to set the stage.

So let's.

Hey, Sam, man. I'm so excited to have you here on the data canteen. So for anybody who doesn't know, Sam was the winner of our 2021 VDSL scholarship. Our very first scholarship. It was to the fourth brain, uh, boot camp for aspiring machine learning engineers. He finished that up a little bit earlier this year.

So I know any of you who are familiar with that will be excited. Hear a little bit to hear about his experience with the fourth brain bootcamp. Um, and he also happens to have some insight on our 2022 scholarships. So that'll be cool to talk about as well, but before we get into all that, like myself and so many other people in this community, Sam has an amazing story to tell about everything that has led him to this point in his life of being interested in data science, machine learning.

Sam, as I was looking over your background, I noted that you were a Navy pilot you've completed award-winning research for satellite design. Your interest includes space flight. You've got a bachelor's degree in airspace engineering. Your hobbies or things like offshore, sailing, cooking, scuba diving, and coffee.

And then of course you're interested in data science machinery. You're like the modern Renaissance man. So, with that, summary of what I gathered from your background, I can't wait to hear the story about how you, wherever you got started grade school, college, military, and then, uh, all the way to where you are now.

[00:02:17] Sam Sipe: Sure. Well, um, thanks, Ted. And, and honestly I have to say it's, it's an absolute pleasure to be here. Uh, I think I told you right after I got the scholarship that, uh, uh, I really wanted to be on the podcast, so I'm, I'm, I'm glad we're making this happen. Um, one thing I'll start off with is, is, um, you know, your journey is your own and, and, you know, I remember being a kid and, and having all these really set and defined goals.

That I just knew I was gonna accomplish with no roadblocks. Uh, but that's just not how things work, uh, in life. And it's best to kind of just roll with the punches and go with the flow. So, um, use the term Renaissance man, but I think, uh, master of none is probably what I would use so, um, but I'll, I'll dive right in.

So, uh, I grew up, uh, in the Midwest I'm from Indianapolis, Indiana, um, grew up in a non-military family and, uh, I was however, at, at a young age, obsessed with flying obsessed with space. And it was just, um, I don't really know if it was just kind of the tail end of the shuttle program and kind of America's interest in, in space flight.

I was kind of waning at that point, but still very much alive in the, in the early nineties. Um, but being a pilot for me was kind of the only option for, you know, I, I didn't really apply to a lot of other colleges I applied to, to the Naval academy, I think to Michigan and like two others. Um, somehow I convinced the Senator and a Congressman to, uh, to give me a nomination to get in there.

And, um, you know, really had a, had a great four years, uh, there and, and then, and gotten to fly. But, uh, as a kid, I, I used to, um, I used to spend almost all my time, uh, tinkering with stuff. I was kind of a, a maker and a tinker at heart. Um, and, uh, later in college, I, after getting my first C at the Naval academy and going up to the professor and asking, you know, why, why didn't you know, why didn't I do well?

Like, what am I missing guaranteed first engineering class, everyone doesn't do well. So I didn't know that yet. Um, and he's like, you know, Sam, I think, I think the problem is that you think, you think like an inventor, you don't think like an engineer. And I think he meant that, uh, as somewhat of an insult, um, or not an insult, but rather like, Hey, you need to get better at this.

Um, But to be honest with you, I've, I've kind of taken that and run with it and turned it into part of my, uh, my personality and my way of, of, of building things and, and always questioning things at first principles. So even as a kid, uh, for me, um, I had a twin brother, so I kind of had a pre-installed, uh, best friend and, uh, you know, someone who's always down to tinker on something or take apart a car.

So we spent Mo most of our, uh, most of our childhood building rockets, um, launching those rockets and then getting them stuck in other people's, you know, trees and property. Um, and then a lot of time, you know, building planes and, and kind of getting into some of the early RC planes and stuff like that.

So, um, I was not, I had no interest in, in software for me that wasn't something that I really wanted to learn about. I just wanted to learn about planes and aerospace engineering. Um, You know, went to, went to high school, kind of just, uh, set myself up as best as possible to, to have a shot at going to the Naval academy.

While I always knew that I, I wanted to join the Navy and go to the academy, um, you know, I didn't really start taking it seriously until senior year. Uh, however I was doing well in school and I had a lot of, uh, activity on sports teams and leadership and stuff like that. So, uh, I was able to kind of, uh, get my stuff together and then apply.

Um, shift off to plebe summer, uh, the Naval academy, which for me, uh, someone who didn't quite understand that, that, that flying for the Navy also meant that you had to be in the military. It's just something that from a non-military family, I, I didn't quite connect those dots till I showed up on I day.

And then you have people yelling in your face and you're kind of immediately getting acclimated to, to that journey. So, um, what I think in the long run ended up being healthy, uh, to always kind of question leadership and decisions and, and question your surroundings, uh, which is a great data science, um, uh, you know, a personality to have on board, but, uh, you know, for the military did take some adjustment for me.

So after a brief period of kind of getting used to a new lifestyle, uh, you know, I, I found the aerospace engineering department at the Naval academy, uh, kind of simultaneously as I found the, uh, the offshore sailing team. So I kind of threw my myself head first into both of those. Um, very different, uh, parts of, of the Naval academy experience.

Um, and so I got to do a lot of cool offshore sailing. Uh, for my four years there got to sail to Bermuda, got to go to, um, you know, all the way up to Canada and kind of all over the place, doing cool racing on very fast boats. Um, and you know, and I still to this day do that and it's a lifelong sport and it's one that I see a lot of cool applications for machine learning and data science, cuz it's something that hasn't kind of hit that, uh, racing community yet.

So, uh, kind of have an idea to work on some stuff in my free time around that, that space. But, uh, at the same time, uh, you know, kind of really started focusing heavily on engineering. Um, I never really liked the, the class structure. I, I always wanted to kind of go learn on different stuff. uh, I discovered my sophomore year.

So, uh, we called that youngster year at the Naval academy that I could do research and actually get credit for it. So once, once I started going down that, um, that road, uh, I finally convinced some, some folks at the Naval academy to let me build and launch a satellite. Um, so I spent the next three years, uh, building a team, uh, of which I was really lucky to have some awesome, awesome folks on that team, uh, helped me build and launch that satellite.

Um, and it went up in 2015, uh, when I was in flight school. So that was a, a pretty cool accomplishment, uh, even after I graduated. So other than that, you know, I, I studied aerospace engineering. I, I did a little bit of coding and software in college, but mostly MATLAB. Uh, and to be honest with you, it wasn't something that I was super interested in.

Um, You know, at the time I was, I was very interested in kind of the orbit orbital dynamics, as well as, you know, satellite constellation, mission design, all that kind of stuff. Um, and it's, it's a really, really big field, uh, that you can kind of dive head first into without writing a single line of code.

Um, but I did realize, and, and one of my key takeaways from, um, from the satellite project and, and delivering it is that, you know, any hardware project has some element of software in it. There is no pure hardware project. So I knew kind of leaving the Naval academy that if I wanted to compete in the modern job landscape, you know, let's say after flying is, is complete, or even during that time that, uh, writing software was, was gonna be part of that.

So in my free time and, and in, you know, in my own investigations, um, While I was in flight school whenever, you know, it wasn't super busy. Uh, for whatever reason, I was always learning more about software learning about Python, JavaScript. Um, just so that I knew when I had to turn the corner at some point in my career that, uh, I was gonna be, um, you know, a generalist or, or be able to do, do the software roles, but also understand the, the hardware and the, the dynamics as well.

Um, so once I, once I graduated, uh, from the Naval academy and commissioned into the Navy, uh, through myself, uh, wholeheartedly into flight school, um, which is a really cool experience, it's a, it's a, um, it's one of the few times I think, or one of the few places in the military where you really get to just learn and be dedicated to a task 24 7, and then your only job really in the, in the whole world for two years is to just learn to fly.

Um, so I really enjoyed that. Um, and I really love flying and, and, uh, I thought it was just like one of the coolest things that you can just take the system, learn everything there was to learn about that system and then go strap it on and, and do it. So got to fly, uh, four different aircraft and then was assigned a fleet aircraft to the E two D uh, which is the advanced Hawkey for those, uh, who, who don't have this memorized like I do.

Um, and then went to my squadron and, and really enjoyed it there. Um, however, I kind of learned pretty, pretty early on as I kind of insinuated that, um, the Navy life long term probably wasn't for me. And, um, a after spending some, some, doing some deep thought and, and realizing, you know, Hey, the odds of going to NASA or going CPS then going to NASA nowadays are, are just not as high as they used to be in the mid nineties or late eighties.

Uh, and, um, so I kind of decided that. I was going to really advance my career. I, I was gonna have to leave the military. Um, and it was a really hard decision cause I love flying. Um, but I kind of knew, uh, it was, it was time, time to go. Uh, really by the time I, I set foot in my first squadron. So, um, anyway, I got out of the, uh, I got outta the military at the worst possible time, right at the beginning of COVID.

Uh, I, I remember being so excited when I, I got my orders to get out and then I knew, um, you know, you kind of don't get to pick the date. The dates picked for you and then world events happen. And I was like, oh, that's this is COVID things gonna be interesting. I was like, well, I'm changing jobs here in a few months.

This is gonna be interesting. So, um, it, it actually in hindsight ended up being a blessing, cuz I think sometimes taking challenges, head on can make you think a little differently and um, So rather than my, my planned endeavor, which was to just travel for a few months and take a ton of time off, uh, I ended up moving to key west Florida and living on my sailboat for a few months, uh, with my then girlfriend and now fiance.

I immediately, uh, started working at a small company called Aaron technologies, which is a, a government contracting consulting company that works mostly with, uh, with satellite satellite data and satellite mission design for some of our intelligence community partners. Um, and oddly enough, I actually got into that job.

Uh, I wasn't really looking for a job at the time. I was trying to just decompress and, and chill after, after leaving the Navy. Uh, but I got a call from my research advisor. Who's still a really great friend and mentor to me to this day. Um, from the Naval academy, he's a research advisor on the satellite that we built.

Uh, and he said, Hey, there's this job opening at this company? Um, I think you might really like it. I think you should take the call. I know you're not really looking yet, but I think you might, uh, might like it. And, uh, pretty much got hired on the phone, started immediately, uh, working for them. And, uh, two years later I I'm actually still working for them.

Um, and it's been cool to kind of get thrown right off the deep end into some really, really tough problems, uh, across the intelligence community research effort. Um, and that, that pretty much, uh, brings us here, Ted. So, you know, one of the things I, I learned right off the bat, uh, like I said before, uh, that software is gonna be critical.

And in, in the same way, I feel AI and ML is critical for the advancement of that software, uh, into the modern age. So, uh, I knew early on, um, that learning more about machine learning and data science was gonna pay dividends. Um, so I initially became interested in fourth brain. Uh, which we'll get into here in a bit, um, probably a year before the course, uh, right, right.

When I first came out, I think someone sent me a link to it and I was like, Hey, this is interesting if I get the time. Uh, and I was able to, to kind of fit it into my schedule, uh, and really, really learned a lot. So anyway, that brings us to, uh, to right now.

[00:14:46] Ted Hallum: Cool. Okay. So reflecting on what you just laid out for us, it sounds like you started out with a passion and a, uh, and a significant interest in flying satellites space, all that kind of thing that led you to you grew up, you ended up going to the NA Naval academy where you studied engineering, um, that led to an interest in software and software design, because you saw the need for that in relation to engineering.

And while you were there at the Naval academy, you got to design and launch a satellite. And then after you finished up at the Naval academy, You were able to realize that dream and become a pilot and actually do quite a bit of flying. Um, then you got out and you kind of picked up with that engineering interest as a space systems engineer.

Um, and then, like you said, you just finished up the fourth brain program where it sounds like in addition to software engineering, you hope to add on the abilities of AI machine learning. Um, so that, that's just fascinating. I have to ask. I remember in high school, I really liked the movie October sky. Is that a movie that, um, is a favorite of yours?

[00:15:54] Sam Sipe: Absolutely. I think I watch it at least once a year. It's a really phenomenal film, uh, for anyone that hasn't seen, it's, it's worth the, you know, finding it, I won the streaming platforms, but, uh, I think I had a, a poster of that on my wall, uh, in high, in high school. I was very much inspired by, uh, some of the paths, people like Homer Hickum and stuff like that.

People that, that, um, similar to me, they, you know, they had big dreams and, and maybe they didn't meet them exactly. But they, they kind of pushed along the way and, and, um, you know, at least successfully made a career out of it. I think one of the things that's, that's hardest, especially for people in the military where, um, very often, uh, you could join the military, looking to do one thing, and then it doesn't work out for one reason or another.

Uh, which seems to be the story that I hear more often than not from friends and, and, and folks in the military. So it's a very common, uh, path to kind of roll with the punches and, and, you know, be nimble and, and learn across different fields and, and sectors that. So you can kind of be a generalist and apply

[00:17:00] Ted Hallum: yourself.

Absolutely now, you know, we've had a, a number of people come on the show now who have a background in engineering. It seems that the training and the background, um, and the fundamentals that engineers have and bring to the table tends to give them a, a significant advantage over a lot of other people, um, in the space of AI machine learning.

Um, do you agree with that? I guess, first of all, and then if so, what do you think some of those fundamental benefits are for people with an engineering background?

[00:17:37] Sam Sipe: Yeah, so engineering is, is less of a profession and it's more of a mindset. Um, and it, and it really is interdisciplinary.

Like, just because you studied aerospace engineer, doesn't mean that you couldn't learn some of the fundamental aspects of mechanical engineering and be just as successful. In fact, between those two, for instance, there's a lot of fundamental courses that you'd take between both. Um, for me, I, I picked engineering, um, at a very early age, I knew it was gonna be an aerospace engineer.

It didn't really matter, uh, for me, but I was very sure that engineering was the right choice. When I, I realized that it was a way of thinking, uh, now I already told you that I wasn't very good initially at that way of thinking. And I was told that I, I, I needed to learn it better. Uh, and it was hard. It, it, and it is hard to kind of, um, tear things down to first principles or, or, or tear things down to systems.

Um, but honestly that skill, uh, led me to the success that I had in flight school. And this same skill has led me to success as a space systems engineer, and as a machine learning engineer. So being able to, to tear apart ideas or systems and, and put them into bins in your head and then connect those systems.

um, is very crucial and it sounds kind of ethereal, uh, and like hard to understand, but, um, you know, I, if someone is interested in engineering, I would absolutely recommend that they pursue it, even if they're not necessarily interested in, in becoming an engineer, it's much harder to do it the other way around.

[00:19:06] Ted Hallum: So for those folks that haven't heard about it, could you tell 'em a little bit about, what fourth brain is, what the boot camp is like, and what sort of graduates did they try to produce?

and, and what they prepare you for, like in terms of the workforce,

[00:19:22] Sam Sipe: Yeah of of course. So one thing I will, um, caveat is, you know, fourth brain is a bootcamp in the sense that it's taking you from some lack of knowledge, to some gain of knowledge, um, on a pathway. But, uh, one thing I think that's a little different about fourth brain is there's a higher expectation.

Um, at least compared to some other bootcamps that I obviously have not participated in. So this is an opinion. Um, you need to have a little bit more knowledge coming into it and more curiosity versus just, Hey, I know I'm gonna take this JavaScript bootcamp, cuz I know nothing about JavaScript and I want come out the back end applying for a, I don't know, front end engineering software engineering role, uh, with fourth brain, you, you, you should have some either self-taught knowledge of, of at least Python.

Um, And some fundamental understanding of Linux and, and some things like that. But, um, success is, is much easier. If you are already a data scientist or software engineer, I was none of those things. So I'm here to tell you, you can do it, but, um, it is, it is. I think one of the criticisms that some folks have said about fourth brain is that the, the barrier to entry is high just on the knowledge front.

Um, but I think with machine learning, it's pretty hard to remove that. Cuz machine learning is itself a very hard concept. It's very advanced, uh, linear algebra, vector chain, rule, gradient, descent, all this stuff that, um, if you didn't have a background in those things, um, it's just gonna be harder for you to, to pick it up.

So having been an engineer, having take, taken a lot of calculus courses, Um, though I personally did not do great in them. Um, and at least understanding these fundamentals, uh, is gonna kind of ensure success. So if someone is, I guess, to boil that down, uh, to those listening, you know, if, if you just got out of the military and maybe you weren't as technical, kind of like me, you weren't, as you weren't running code software while you're in the military, um, might recommend taking, you know, a Python bootcamp or something like that first.

Um, and then pivoting that knowledge into something like fourth brain. Sure.

[00:21:39] Ted Hallum: So you would say fourth brain is not really the bootcamp where you would go to get started with data wrangling. You probably should be pretty comfortable with that. Like you say, comfortable with Python. Um, and then this is more of an intermediate to advanced level learning experience.

[00:21:56] Sam Sipe: exactly. And, and just to talk a little bit more about, uh, my personal, um, Experience. So I, I had already been working for about a year on a large, uh, government contract doing machine learning, uh, before I took fourth brain. So I kind of had a fundamental understanding of the objectives of machine learning.

I knew, uh, you know, how some of the basic pie torch intensive flow models worked. I had played around with open CV. I had done some other, um, not boot camp, but like other courses online, like many courses like, uh, pie image church is a great one for anyone looking for it. Uh, just some just to see if they're interested in machine learning.

Um, and that really gave me a leg up, one thing I would say to. uh, who's interested in ultimately getting into machine learning is, is really getting into data and data science first. Uh, and that's why it's veterans in data science and machine learning, not just veterans in machine learning, cuz machine learning is, is a subset.

Um, and it is hard to go straight off the deep end. So rooting yourself in data and data science, whether it's data engineering, data analytics, business intelligence, any of anything in that field. Um, and then going into machine learning is, is a much more, uh, healthy and achievable approach than just saying, you know, Hey, I just got out of the air force.

I was an avionics tech and now I wanna go be a machine learning engineer. So, um, you can do that, but you're just gonna have to build some fundamental understanding first.

[00:23:25] Ted Hallum: Absolutely. So you had the advantage of getting a good foundation there in a year, uh, at your current company doing machine learning before you started the fourth brain bootcamp.

Um, now. That bootcamp is a intense experience that spans multiple months. I know you learned a ton going through that, but if you were gonna distill that down and think about maybe the, the top two or three concepts, valuable takeaways, or abilities that you walked away with, what would you say those are?

[00:23:58] Sam Sipe: So my, my goal going into, so I had a, a very clearly defined goal. Uh, and I think I talked about this Ted during our, our interview for, for me getting a scholarship. But my goal for fourth brain was to personally discover when it is necessary to use machine learning, either classical or some of the newer, um, deep neural network, machine learning, one it's best to use machine learning and one it's best to solve your problem else, you know, with heuristics or in a different way.

Um, so, so that was, that was something I feel like I, I did achieve, um, And I did discover that it wasn't just a binary question. It was, it was, it was a, um, a gamut of, of things you might select. So, um, I now have kind of a, a playbook of, Hey, here's your data? Um, how would you approach that data? What would you do if you knew that you had already tried some other approaches and now need to use machine learning, what would you do if those classical machine learnings, uh, techniques aren't working, um, and stuff like that.

So, uh, really fourth brain is, is really good at showing you a wide gamut of machine learning approaches from your unsupervised semi-supervised fully supervised. Uh, and then some of the stuff on the, on the cutting edge, like reinforcement learning at a very high level. Um, but, but truthfully it's, it's exposing you to kind of the playbook and, and the options.

And then. Getting you to that. Uh, you know, I always say the 40, 70 rule, so knowing 40% about something now you can make a decision about it, but you certainly don't know more than 70% and you're certainly not an expert. So it kind of gets you in that sweet spot, 40 to 70% of the knowledge on, on something so that you could go learn more.

Cause machine learning is, is one of those things that it's hard to just sit down and, uh, and watch YouTube videos or take Corsera courses like you really have to dive in and you're gonna have to ask questions that are, are hard to understand, and that require, um, discussions

[00:25:57] Ted Hallum: mm-hmm yeah. Now talking about the challenge of trying to learn this stuff from like YouTube videos or something like that, anyone who's done significant upskilling in this space over a period of time, they've probably come to the realization that the best way to learn this kind of subject matter is through projects.

Um, because you get into a real world project and you run into challenges and hurdles and you have to overcome them. And that's where. You know, even if you've heard this stuff before, that's where the knowledge really takes root and becomes meaningful and you set it to memory and it becomes something that you can reach back into your mind and quickly make use of, you know, on your job or whatever.

So I think one of the things that's really cool about fourth Green's bootcamp is I know that they have pretty intense capstone that, um, that marks the end of the program for graduates and having done some virtual coffees with you. I know that you were able to work on a really cool capstone project. So, uh, I'd love if you would sort of launch into that, tell us, you know, what the name of that project is, what did it aspire to do?

Um, sort of from the vision standpoint, and then we'll go from there.

[00:27:08] Sam Sipe: Absolutely. So, um, the thing, so I will initially say, and I should have said this at the forefront, that the thing that really attracted me to fourth brain was the project aspect, because. If you haven't figured out by now, I'm the type of person that learns by building something, not by just listening to, uh, YouTube videos or, or an instructor.

So, uh, half a fourth brain, uh, the second half is all a capstone project. Um, and actually there, there is some project element, like a mini project in the first half, two that, uh, can help you learn. Um, and, and our project was called amplify. Uh, it was a team of, uh, three folks. So two other folks and me, um, two other really, really capable engineers, uh, Christian welling and, and John Cher here out there, thanks for all the hard work.

And, uh, you guys should hire them. They're amazing. Um, so the three of us worked on, uh, amplify, and if you wanna check it out, it should still be live at amplifygrid.com. Um, but it's a deployed web application that does, uh, it's a deployed machine learning workflow that does end to end machine learning. Which is kind of the, uh, the pinnacle in the machine learning community of taking something from a concept or just exploratory data analysis, or just a data source to the actual model development and training the model deployment, and then the inference.

So it's, it's a whole pipeline of stuff that happens, uh, all in one. So, so what is it? So, um, you know, despite being also interested in space and flying and drones and all the other things that, um, that you and I have mentioned and talked about, uh, I'm also really passionate about energy in the climate. Uh, I just see it as a large, um, a large and obvious problem in our society.

And, uh, I think it's important that we spend a lot of time as a society finding and identifying people to go solve these problems. um, and not so we can just Terraform the earth, but, but so we can control how people use energy and things like that. So one of the things that comes down to is, is people being even exposed or knowing what sort of energy they use and, um, and kind of monitoring that and predicting that so amplify, which you should see if you've gone to the website, find out amplify grid.com, uh, predicts power usage and generation from solar panels, 48 hours in advance.

And it, it does this, uh, mostly by using weather data and some other, uh, data sources, namely the position of the sun, the elevation of the sun, stuff like that. Um, and using a, uh, a, a L S TM model, a long short term, uh, memory model. We were able to, to do that and with a pretty remarkable amount of accuracy.

So within four kilowatt hours, we were able, or four kilowatts, excuse me, Um, 48 hours in advance, which is pretty cool. Um, now guaranteed, this was, uh, all set up specifically to a building actually that's right outside by window that has solar panels and, um, a lot of sensors on all of the lines, uh, power lines.

So they know exactly how much power each apartment, how much power the solar panels are generating. Uh, so it's very tailored to that specific apartment building. Um, but it could be, you know, deployed and, and commercialized across different, uh, different aspects. I, if necessary, um, we didn't really build it to solve like a business problem necessarily, but rather a machine learning problem.

And so that's another thing I've learned, uh, through fourth brain is that a lot of times, um, you can do really cool stuff with machine learning. Uh, but sometimes you have to, to step back and ask, are you asking the right question? Like, are you solving the right problem? Do do your metrics make sense? Uh, what is your success criteria?

So one of the things we learned with amplify is, yes, it's very cool to predict how much power you're gonna use into the future. Uh, but like who cares? Right? So one of the things we added to kind of answer that question was, um, folks, uh, so at this building next to me, there are, um, 13 electric vehicle chargers booked up.

And so one of the things we thought is, well, that's a pretty big load. It would probably be best to optimize, you know, when that load should occur or when, if you had to charge 10 cars when you should charge them. Um, so we didn't actually control anything on the building, but we did write a script, uh, with some secret sauce that John came up with.

Uh, it was just brilliant to, to figure out when, to charge those vehicles. So if you see on the site, there's, um, you know, there's, there's little green bars that tell you when the best time to charge an electric vehicle is now it turns out that, um, Obviously, if power's cheap in the middle of the night time of use rates, that's the best time to do it.

But, uh, what is kind of cool is you can see those thresholds when the predicted generation starts to go enough above the predicted usage. It'll tell you to charge in the middle of the day, too. So in the future, when everyone has an electric car and can't everyone can't just charge when they get home at 5:00 PM, we're gonna have to schedule these things.

So there are some business uses, um, coming down the line. Uh, but for the most part, I think, um, for me, it was pretty eye opening to see, uh, how you can take a very, very messy data source so that the data sources that we are using were very hard to use. And we spent at least 80% of our time just cleaning data.

Which is kind of crazy when you think about it, cuz I just told you it's a machine learning course, but realistically, if you go out there like we did and you find your own data source, you're gonna spend a lot of time debugging, you know, one, we had issues with look ahead air. We had issues with time zone.

We had issues with, um, data that was just missing for some reason, like the something rebooted and data didn't show up. And when you have like thousands and thousands and thousands of lines of data, you can't just go through it. So it has to be automated. Um, so if you're out there and you're thinking about doing fourth brain and you write Python, you know, Python, you know, knowing pandas and NPY, uh, num pie are really, really gonna help you be successful at fourth brain.

And, um, honestly I think we would've been, we would've been in trouble without, uh, Christian and John's knowledge, uh, of, of those two, uh, software packages and being able to make use of that before we even started any model model development are the real. Sensor flow and machine learning. Um, but what was also cool is, you know, I had a background in cloud engineering, cloud deployment.

So, uh, once, once we got it all built, uh, and Christian had some front end experience, like the team really came together to, to do, um, to do some work, to, to build out the front end and, and make it work. So the one thing I'll say is that machine learning is not something that just exists in isolation anymore.

So, uh, I think the best machine learning engineers are the machine learning engineers that are also data scientists that are also software engineers. Um, and that don't just build models kind of on their own, but integrate with other teams and integrate frankly, into the DevOps and ML op cycles that you see about now.

[00:34:28] Ted Hallum: Absolutely. So, um, I love that you guys. Pick the difficult issue, a challenging issue. You picked a relevant issue to our time. Something that's just gonna become increasingly more important, as you said. Um, as more and more people get electric vehicles, uh, they consume a tremendous amount of energy, uh, just to put in perspective, uh, the, the Tesla cyber truck that's expected to come out next year.

I've heard that if you get the wall charging station for your house, so it can, um, or the wall power station rather for your house. So it can act as like a backup power spa. If your power goes out, mm-hmm , that truck will be able to power your house to include like your HVAC and everything for two weeks.

That's the estimate. So that gives some sense of how much power is stored in these vehicles. And if they're, um, you know, if the battery's significantly depleted and you're doing a full recharge, I mean, that's a tremendous pull on electricity. And once that becomes normal, as you said, and everyone has that type of vehicle, then a system like the one that you guys worked on there in your capstone.

Is, it's not gonna be a nice to have, it's gonna be something that, uh, utility companies absolutely have to have and roll out, um, across, you know, residential buildings and commercial buildings everywhere, um, so that they can, you can manage that load. And everybody's not trying to get the same thing at the same time.

Um, also appreciate you being transparent about some of the challenges that you guys ran into. You mentioned, um, you know, what we all realized with real real world projects and world data is that it's always a tragic disaster that you have to somehow clean up and make usable. Um, aside from the dirty data, were there any other notable hurdles that you guys ran into that you had to overcome?

I know sometimes simply data availability, like trying to get the right data for the project you wanna do is, is a challenge. So anything along those lines or otherwise that you haven't mention.

[00:36:23] Sam Sipe: Yeah. I mean, there's a huge list. Um, so keep in mind, this all happened in eight weeks and I think if we could go back, like we bid off a lot, you know, and our, our project was very complicated and for us success criteria was getting to the finish line because it was just a lot.

Um, but, but one thing to keep in mind, yeah. With this machine learning project, the model is a very small part of it. And people spend a lot of time at universities and, and research labs and other places developing these amazing models. Um, and they're good. So really, uh, and you hear about it all the time now, but data centric, AI and ML is the future because trying to reinvent how an LSDM works, uh, during your fourth brain capstone is just, you're just gonna pull your hair out.

So, um, so the clean data is, is the hardest part. So to give you an example of that, the, um, We had all these issues with normalization. So normalization allows you to kind of scale all the features in your data set. So features in this case being power usage or position of the sun or day of the week. Um, and one thing we found is that because the week is the soft tooth, right?

So it's one, then it's two, then it's three. Uh, we are able to kind of, rather than doing that, we were able to sign you Soly and code the day of the week and the position of the sun and turn it into two features that were basically revolving around a, a circle. So using sign cosign. So like I said, deeply, deeply, deeply rooted in math.

Um, and actually like once we turned that on, it just started working. It was really cool. Um, and I think, uh, I think that was a, a John and Christian 3:00 AM revelation after finding someone's paper on it. And we're like, we could do this. That's not that hard. So, um, And honestly, there's little 3:00 AM revelations when you're exhausted and you just figure out something, uh, they make this stuff addicting.

Like they make ML really fun. Um, and, and what's nice is it's the code is actually, once you get good at it, the code's pretty easy to write a lot of API integration, uh, in interaction with Python. Um, but like when you can really kind of see how the model's working and how the data's working, it makes you feel, uh, like you're learning a lot and that, that it's, that's actually doing the thing that you built it to do.

Um, so to your previous comment though, you know, about the vehicle to grid or battery storage, uh, that that's really kind of what inspired us initially to kind of go down this rabbit hole, um, again, building something that isn't quite necessary yet. Uh, so one of the things we see amplify being like useful for is say, you.

13, cyber trucks all hooked up out there and there's a power outage. You could, you know, use the, this controller to control the microgrid, uh, that now exists between the trucks and, and the apartments. Um, or you could figure out when to best charge and discharge those trucks, or in my case, you know, discharge my car or even just batteries that are installed in the basement.

Right. You could do that too. So, um, but to give listeners kind of those who aren't aware of the loads required to charge electric vehicles. Um, you know, my car is 80 kilowatt, uh, 80 kilo kilowatt hours of energy stored in it. Um, so when you get home, you know, with a low state of charge and you charge it all the way up, that's 80 K that's like, you know, an average American household uses like 16 kilowatt hours in a day.

And you just use that in a couple hours. So, uh, it can be a lot of power all at once. Um, so it, it, it is, it is a problem that we will have very soon.

[00:40:05] Ted Hallum: Cool. So, um, changing gears a little bit. So, and last year you were the beneficiary of VDS mill scholarship. Having done the bootcamp and had a good experience.

Now you're very generously. You've come back and you're actually serving on our scholarship committee for 2022. So you've got some cool insight about what we're hoping to do with scholarships, uh, for the coming year. And those will be rolling out in just a couple months later this summer. Um, so if you could just introduce, um, the audience who might be interested in scholarships to what you guys have in mind for, uh, 20, 22.

[00:40:46] Sam Sipe: Yeah, absolutely. So I'll start off with, you know, to the. P D SML members listening. Thank you seriously. Uh, literally couldn't have done it without you, so thanks for your support, uh, both voluntarily and, and otherwise. Um, and, and for those of you that aren't, uh, members yet and are, think about joining, uh, please do so.

Cause it's a great community. Um, so you know, one of the things, uh, I said I wanted to do is give back, uh, you know, I thought this, the scholarship is such a great idea and it enables people that are ML tangential or AI ML, data science, tangential to really break out into the community and, and to be successful.

Um, and frankly, you know, having, having been picked for a scholarship makes you, you know, kind of cuts through some of that imposter syndrome that we all suffer as, uh, as folks leaving the military, cuz you just feel like you're disconnected from something, you know, spent six years, 10 years, whatever, doing something, not on the critical path of the tech world or the AI world, you feel separate.

Um, so, you know, having the support of other veterans and stuff like that, monetarily and otherwise, uh, really kind of makes you feel like you can be successful. So for me, it was critical in, in feeling validated and feeling like I could do this. So thank you. Um, so the, the people that we're looking for, um, well, I'll kind of talk a little bit about, so, you know, um, how the scholarship committee of which I'm now just a member, um, is architecting scholarships of the future.

So one thing we said at the beginning of this is that, uh, hopping right into fourth brain might be a little bit too much. So, um, the scholarship committee has come up with this idea of, of nano scholarships where, uh, we haven't quite figured out how it's gonna be implemented, but, um, we're gonna allow folks to, to do small, uh, courses on their own at self pace courses on things like data camp, and, uh, course era.

Uh, where they can learn Python or they can learn SQL or something like that. Something some useful data science need and really decide if, if they're headed in the right direction. Um, because you know, like I said, it's best to kind of wander around, figure out what you're interested in, uh, first and really be sure yourself rather than get, you know, roped into a fourth brain course on machine learning and realize that you, you know, maybe you didn't really wanna do it in the first place.

Um, I, I will say, you know, during fourth brain that we did have people leave the cohort, it, it, it wasn't like a rubber stamp, easy everyone graduates. Like there were folks that, that didn't make it through. Um, and it, it is, it is challenging. So I think having a mini scholarship, um, will, will help people learn some of those fundamental skills, uh, before they kind of go off and do something as big as fourth brain.

Um, we're also still gonna have the fourth brain scholarship, which is an annual scholarship. Um, That allows, you know, fourth or, uh, BD SML to basically support one person every year, going through, uh, up to a certain percentage, uh, going through fourth brain and it's, uh, you can apply for it now, uh, before you've applied for, for fourth brain.

And then, uh, you know, once you get accepted, then apply to fourth brain and you don't necessarily, you can do it in the other order and you don't necessarily have to do it, uh, immediately. So what I'm trying to say is it's a, um, it is a low burden scholarship. You don't have to write a bunch of essays or anything like that to your Congressman.

It's, it's a very simple, uh, thing. And, and we, we have it to support members to do kind of what I did and to feel validated and supported. Um, there will be, uh, some more information coming on that, on the, uh, PSM L website. So, uh, I recommend that, uh, you go there to look for more information, but, uh, for those listening that maybe don't wanna go that route.

I definitely recommend, um, You check out the, the nano scholarships as well. Uh, cuz it's a great way to start learning, uh, without, you know, being stuck with uh, 16 weeks of, um, of machine learning. It's a, it's a lot. So, um, one thing I will say on that topic kind of as a tangent for you, Ted, that, uh, I mentioned earlier is that, uh, fourth brain is, is a large time commitment.

Um, I think they say it's about 15 to 20 hours. Um, I think that number is, is predicated on someone who is already a data scientist. So if you're someone like me that was kind of ML adjacent or uh, data science adjacent, um, who maybe doesn't have as fundamental of an understanding of Python and, and pandas and stuff like that, it's probably gonna take you more time.

It, it doesn't mean you can't succeed. It just means you're gonna take more time. So I found myself spending at the very least 20 hours a week doing it. And then, um, our capstone was very sophisticated, complicated, uh, although it is boiled down nicely into a neat, neat little web app. So it looks pretty simple from the outside, but, um, when we were doing that, there was many late nights.

Um, so it, it was very large time commitment. And I will say working full time during that was, um, was an extreme challenge. So thanks, you know, to the Aaron tech team for, for letting me throttle back on some of my commitments over there to focus on, on learning. Um, that was, uh, that was definitely critical to, to being successful there.

[00:46:08] Ted Hallum: Well, Sam, thanks for all the, that information about the upcoming 2222 scholarships. I love the direction that the committee is taking our scholarships, because it sounds like for those folks who might be interested in applying to the fourth grade scholarship, because they're at that intermediate level.

They've got the time necessary to commit, to do a bootcamp. That's that intense. That's still gonna be available, but for people who are, earlier in their learning journey, maybe you're not quite sure if the data sphere is for you. You want a good way to try it on for fit and start dipping your toes in the water.

Um, or maybe you're, you're a little bit, you know, into your learning journey and you want to take things more seriously, but you don't have, you know, due to family commitments, work commitments or whatever, you don't have the latitude to make the full on commitment to fourth brain. Um, you can apply to one of these, uh, nano scholarships that are gonna be offered more frequently throughout the year.

Um, and you can apply that to, uh, an appropriate Corsa course or data camp course. Uh, and, and, and exactly which platform it's gonna be relevant to, I know is still, um, we're still solidifying that on the committee side, but that's gonna be available to people that can't make that full fourth brain commitment.

So it there's gonna be a broader array of, of scholarships available and we'll be able to benefit the learning journey of more of our members, uh, in 2022. And I think that's super exciting. So very cool. Thanks for laying that out for us. Um, now what I'm curious to know about you finished the fourth brain program, which was kind of a turning point.

You mentioned that you were, um, AI ML adjacent before doing fourth brain, and you acquired all those skills during the bootcamp. So now you have them, um, are you gonna use that to be. Better in your current role, or when you look out into the near to midterm future, do you have other aspirations on your career trajectory and you hope to use that fourth reign experience to kind of hit pivot, um, into something a little bit different?

[00:48:12] Sam Sipe: Yeah. So I'll say, you know, even in the, after the first couple weeks of, of fourth brain, I noticed, uh, and you know, there's some confirmation and availability bias at play here, uh, for you, uh, Daniel conman folks. But, you know, uh, having, having that knowledge and like going through this course made you kind of see things already a little bit differently.

Uh, and guaranteed, I was working on a large machine learning research and development program. So I was familiar with some of this, uh, but I started to see the problems that we talked about. Uh, you know, uh, having data data sets that were improperly weighted, uh, and stuff like that, or not using normalization or things like that across all these different projects.

And I was like, oh, okay, I'm learning about this right now. So the problems that these teams are trying to solve are something I now understand. So I will say, uh, for someone that was working in the field, um, I saw an immediate gain in my ability to kind of be a value add to the team, cuz I'm starting to understand, um, some of the deep machine learning problems, uh, of class imbalance or something like that.

Um, and then also kind of towards the end, I was able to, to help these teams that were working on this hard problem solve issues that they had, um, which really felt validating, um, now. Uh, where have I gone since then? So, like I said, um, you know, Aaron tech has been really good to me. It's, it's a phenomenal company and it's one of the few places I think, in the IC that you can get plugged in on some really hard and cool problems and, and be able to be as, uh, interactive and, uh, uh, you know, uh, you know, interactive and present as I was.

So, um, but that said, uh, I already said that I had an interest now currently in energy in the climate. So I am starting a new role at a new company, uh, based in Austin. It's a startup called Arbor and we're gonna be working, uh, with a lot of energy data and we're gonna help people save money on, on their, uh, power bill and help 'em switch their, their power supply from, uh, you.

Dirty energy or coal or something like that to, to clean energy, if that's something that they're interested in. So, uh, it kind of goes back to that fundamental, uh, statement I made earlier where people just knowing how much energy they use every month or having a, more of a touchpoint with it, or even just some small nudges, um, can, can help us, you know, start to solve this problem.

So, well,

[00:50:53] Ted Hallum: congratulations on the new role. That sounds amazing. Um, and I'm just gonna go out on a limb here because you and I haven't talked about this, but, um, in the vs mill community, we always stress how important it is to do projects, do projects on your own, make that investment, have a robust GitHub portfolio to showcase your skills and abilities.

Um, and I, I just have to believe that there's, it can't be a coincidence. The nature of the project that you did at fourth. And the nature of this new role, you just landed. I'm guessing that that project really probably played a instrumental role in you landing this new job. Is that, am I, am I right? Or am I off base with that?

[00:51:33] Sam Sipe: No, I think you're correct. So it's, it's not the same problem. So this, this company's in no way trying to solve the, the amplified problem, but, um, right. But they probably

[00:51:41] Ted Hallum: solve your problem. It is to solve energy related problems. Yeah,

[00:51:46] Sam Sipe: exactly. And, and I tell, you know, I've done a lot of interviews with folks myself and, and when a candidate comes along and they seem really passionate and I noticed they don't have any GitHub or they, you know, they have no commits on anything.

It's like, well, you know, are they really that passionate or are they just passionate in this interview? Cuz they want the job and guaranteed, you know, maybe that's okay. Uh, and I also understand, you know, kind of in, in our community and my community, uh, sometimes you can't put everything on GitHub. I've done a lot of stuff, a lot of cool work and a lot of cool ML stuff that I'd love to just throw out there, but it's um, It's not something I can do.

So even my GitHub profile is not as, um, you know, not as available as, as I think it, it could be, but, you know, I do have a lot of other little personal projects on there. And so when you're going for a job and you have a GitHub profile with code that you wrote on it, it really makes you look three dimensional and it makes you look like someone that's not just a coder in the office.

Uh, and then, you know, you do what else with your free time, but you're someone that's actually passionate interested about this and you're gonna push the boundary and you're not gonna be, um, someone that needs, you know, a lot of oversight and management. So, uh, that's one of the first things I look at when I look for candidates and I'm, I'm sure they did the same thing.

Um, amplify is actually still private on GitHub. Um, though, I, I think John and Christian and I have talked about, uh, releasing that, but there is still some, uh, Some twinkling thought about turning that into something else. So we haven't opened that up to, uh, the broader community yet. Um, and that's also cuz some of the data set we don't wanna share with folks because it's, it's literally from people's apartments and stuff like that.

So there's some ethics that as well. Um,

[00:53:29] Ted Hallum: yeah, very cool. Well, um, it was great earlier hearing about the project you did called amplify and we'll make sure that there are links to that in the show notes so that anybody wants to go out and play around with your web app. We'll be able to do that. Um,

[00:53:45] Sam Sipe: now.

Yeah, and there's a, there is a slide deck on there too. If you want to learn more about how we architect the model, all that stuff is, is available on the slide deck on the web amp as well. Oh

[00:53:54] Ted Hallum: fantastic. So you did fourth brain. That was an intense learning experience. Um, and I'm sure you took a little bit of a, a, of a well deserved break after that, but.

Any of us in this field, it's just comes with the territory. The learning is never done. Um, we'll learn the rest of our lives and still only know the tip of the iceberg. So with that being said, I have to ask what's your new, current learning focus?

[00:54:23] Sam Sipe: So, uh, kind of ironically, uh, SQL is one of the things that I wanna learn more about.

Um, it's something that I just haven't had to use it, so I've never had to learn it. I've, I've, I've used some very similar things with similar syntax, but haven't quite gone through SQL. So it's one of the courses I'm, I'm starting on data camp, uh, just to kind of freshen up before I start, uh, working full time at this new role.

Um, I am also always learning more about cloud, uh, cloud deployments, cloud engineering, all that stuff. It's something. Uh, like I said earlier, being a data scientist who knows machine learning, who knows DevOps, who knows, uh, cloud engineering and all these things makes you a really, really, uh, attractive to employers, even if you're not doing all three of those things in your job, uh, and obviously working in a startup, you do have the ability to do more than just, uh, one thing.

And you get to experience different, uh, stuff across the company. Uh, but it allows you to be, uh, to kind of understand like, and build stuff in a better way so that you can make other people's other developers life easier. Um, so I'm always learning more, uh, about cloud deployments and stuff like that.

Tons of Kubernetes learning lately, learning Terraform, uh, just cuz I've mostly been using AWS. Um, so I'm trying to kind of move away from AWS and learn more infrastructures code elsewhere. Um, doing a lot of Google cloud stuff recently, which is cool. Uh, so, you know, kind of seeing how stuff is built over there.

Which is very different. Um, and then I'm kind of tinkering around with some more off the shelf ML stuff, um, so that you can do machine learning. Uh, you can solve the same machine learning problems without having to build your own algorithm, but, uh, you know, use, use some of the managed services that AWS and Google and Azure are now offering well real quick.

[00:56:16] Ted Hallum: So you mentioned, um, you having used AWS and Google cloud platform, and then there, the very end you mentioned Azure. I'm not sure if you've actually used that one as well. Um, but for people who've maybe only used one or maybe they've just never really touched cloud computing, um, real quickly, how would you kind of rack and stack those?

Do you have a favorite? Um, are, are, what are the main differentiators between the, the three major players there.

[00:56:45] Sam Sipe: So really, um, you know, I, I've spent a lot of time thinking about this and working on this, uh, mostly with government customers and I've spent almost all of my own time working with AWS, just cuz of how contracts are set up.

AWS tends to be, uh, more present, however that is changing quickly. So, uh, right now, if, if you're listening to this and you're, you're not in the data science community, but rather you're a cloud engineer and you only know AWS, you know, learning Google cloud and learning, uh, Azure is probably worth your time because, uh, if they're very similar and pretty much anything you can find across any of the three of those is present somewhere in the other one.

Um, you know, they each kind of have their own specialties. Um, but for the most part, like, you know, if, if you wanna do something on AWS, there's a way to do it on, on Google cloud as well. In terms of ease of entry, if you're someone that hasn't learned any of them and you want to, you wanna start learning?

Uh, I think Google cloud is probably the most approachable. Uh, and, and if you want to try some of their, uh, managed data and ML services, they're probably the easiest to set up. Um, you know, if for enterprise customers and stuff like that, they tend to lean over to the AWS side of the house, cuz things are harder to set up, but more, uh, tuneable uh, out of the box.

Uh, but realistically you can get the same sort of, um, the same cloud. Goodness, if you will, from, from all, from all three of those and there's more coming up behind them, right? You hear about IBM and Oracle all the time and, and Salesforce has a cloud, so there's lots of other clouds you could learn too.

Uh, if you're kind of in that ecosystem. So it doesn't have to be one of those three. Um, and uh, but really, you know, mastering one and, and, and understanding kind of fundamentally how the system works. Uh, it, you can find it's pretty easy to move to other ones, cuz a lot of it's architected the same way in the hyperscalers.

[00:58:42] Ted Hallum: Very cool. Thanks for that insight now. Um, among VDSL members, one of the big benefits of being in a community is we get to share tips and tricks and Hey, here's the latest book I read that really helped me or here's this course is helpful. So I'd love for you to share your favorite learning resources.

Um, I know as a long term BS, Melin, your number one favorite podcast has to be the data canteen, but we'd love to hear about the distant second favorite, um, as well as, uh, your books on machine learning. Um, course here, courses, that kind of thing. Sure.

[00:59:19] Sam Sipe: Well, let's start with podcast cuz you know, Hey we're, that's where we are right now.

Um, so Lex is always good for, for folks for, for ML stuff. Uh, Lex Friedman, great resource, uh, A lot of times you'll go on there listening about something and just learn something completely new. Like I learned the other day that, the bacteria on the space station has evolved to the point where it is a completely separate species.

So there is bacteria that exists on the space station that when astronauts, you know, go orbit and come back is now present in their bodies. That's not available on earth, which is just like, wow. Like after 20 years, plus a little more than 20 years on orbit, um, bacteria have evolved for, for ZG. So that that's cool.

So, Lex, I learned that on Lex. Um, I new one, I just discovered that, that I wanna share that's kind of ML adjacent, but also an energy is catalyst, uh, with Shayla con really cool podcast, uh, very energy specific, but, uh, a lot of kind of crossover with stuff kind of in, uh, in the ML space as well. Uh, in terms of books, um, Thinking fast and slow by Daniel.

Conneman amazing, amazing tone. And you should absolutely pick it up, um, or listen to it. Uh, I, I tend to listen to books. I'm just more of an auditory learner in general. Um, and for long road trips and stuff. It's great to listen to that. Um, I also read, um, Nate silver from 5 38, his, uh, signal in the noise, another really awesome book.

And I was actually reading that during fourth brain, um, which was kind of cool because there's a lot of really good crossovers with what he was doing, like K means clustering and stuff like that for his data, for both baseball and politics. Um, and it was really cool. It kind of helped solidify some learning for me.

So that's, that's always a recommendation on my list. Um, so I'll, I'll leave you with just those two as far as resources, um, just for, for peer, you know, machine learning interest. Um, you know, YouTube is, is always great. I know. Uh, was it the three through brown, one blue? I think we've all spent hours getting sucked down the rabbit hole of some of those YouTube videos.

Uh, they're just fantastic. Um, so anytime I don't understand a concept, uh, specific concept, you can usually find a YouTube video on it. That's just way down the rabbit hole, which is cool. So that's a good one. Uh, pie church is a cool one for, for open C and learning computer vision stuff. Uh, that's a, that's an awesome resource.

Um, you know, fourth, brain's an awesome resource, but it takes a little while to get into it. So that gotta throw that out there. And then, um, you know, from as far as like the paid courses and stuff, uh, data camp seems to be, uh, a really great resource, but it's obviously not free. Um, but really, uh, I think what's, what's great.

Is that searching for some of this stuff now is, is a lot easier. I know a couple years ago, if you having an issue with, um, searching for, I don't know, some class imbalance problem, just since I threw that out earlier, it, you couldn't find an article on it, but now you can search it. And towards data science, you know, medium, uh, article will come up right at the top and someone who spent a lot of time, uh, writing and think of thinking about these problems has, has put, uh, put information out there.

So, uh, what's nice is that it's now easier than ever to get resources on ML and engineering and data science. Um, so I highly recommend that, uh, you, you know, you search hard for it and, and you'll find what you're looking for.

[01:02:58] Ted Hallum: Yeah, absolutely. Sam, we are very fortunate to be doing, uh, data science and machine learning in the era that we're in when, um, we're so flush with resources and such an information rich environment. So with that, I think we're about at the end of our time. Thank you so much for being so generous to come on and tell us about your experience with fourth brain, um, your data gen your data journey more generally, and the direction that our new scholarships are going in 2022.

I think the last thing to cover, uh, throughout the episode here, we've had, um, your, your LinkedIn username. So that's one way that people can reach out to you, but what's your preferred means of contact for people that have watched this, maybe your story inspired them or sparked questions they'd like to reach out.

Great.

[01:03:42] Sam Sipe: Yeah. So, uh, LinkedIn's always an easy, um, easy way to get ahold of me. It's one of the few social medias that I'm, I'm still on. So, uh, that's, that's probably the easiest. If you wanna send a quick message, uh, feel free, um, for whatever reason you're not on that platform, uh, you know, or, or don't wanna sign up, uh, I would say join VD SML, cuz we have a little slack channel going there.

So that's, that's a really great way to come find me. Um, I, I hang out on slack there and, and, and, uh, try to participate as much as I can and help folks out. Um, so, you know, joining VD SML and joining that slack channel is a really easy way to get direct contact. Um, and then, you know, I'm also comfortable with, with folks emailing in.

So if you wanna, if you listen to this and you heard me, um, you know, can go ahead and shoot me an email, it should be pretty defined, pretty easy to find nowadays, but you can grab that from my website, Sam sip.com, or you can go to site.io and it'll take you right there.

[01:04:40] Ted Hallum: Fantastic. All right. Well, again, Sam, thanks much for coming on.

Love talking to you. Look forward to hearing from you again in the future and take care.

[01:04:50] Sam Sipe: Thanks so much, Ted.

[01:04:51] Ted Hallum: Thank you for joining Sam for this conversation about his unique journey to the data sphere and how the VDSL community and our 2021 scholarship have empowered him to pivot into the next chapter of his career. With that until the next episode I bid you clean data, low P values, and God speed on your data journey.