Value Driven Data Science: Boost your impact. Earn what you’re worth. Rewrite your career algorithm.

[00:00:00] Dr Genevieve Hayes: Hello, and welcome to Your Value Boost from Value-Driven Data Science, the podcast that helps data scientists transform their technical expertise into tangible business value, career autonomy, and financial reward. I'm Dr. Genevieve Hayes. And I'm here again with Daniel Bourke, co-creator of the Nutrify app and Machine learning instructor at Zero to Mastery Academy, to turbocharge your data science career in less time than it takes to run a simple query.
[00:00:32] In today's episode, you'll discover practical strategies for validating data product ideas before investing significant development time. Welcome back, Daniel.
[00:00:42] Daniel Bourke: It's great to back. Thank you, Genevieve.
[00:00:45] Dr Genevieve Hayes: Pretty much every independent data professional has a story of the product or service they launched that nobody wanted to buy. In fact, one of the biggest risks for self-employed data professionals is spending months or even years developing something only to launch to crickets. According to your website, Daniel, the machine learning courses you teach at Zero to Mastery Academy have had over 160,000 students worldwide.
[00:01:11] And by the end of 2023, nutrify had over 10,000 paying customers. Now clearly you've managed to create products that people actually want and are willing to pay for. Let's focus on Nutrify. Firstly, can you briefly explain what Nutrify is and what problem it solves?
[00:01:33] Daniel Bourke: So I just wanna correct that the 10,000 paying customers, I wish it was that. That was their goal. So we're probably a hundred x less than that. In terms of paying customers. But I like to just state in words of like, Hey, we want to do this.
[00:01:46] But on the student side, I haven't updated my website, so that's probably closer to 250,000 now. So that's a little bit higher. One's lower, one's higher. So Nutrify is a food education app. My background is food science and nutrition. And then I got into machine learning and naturally with computer vision.
[00:02:05] When I learned about that I was like, well as a nutritionist, I want to see if I can embody if you take a photo of something and it give the nutrition information back from that. And so that's the main premise of nutrify. You take a photo of food, it will tell you about the nutrition that's in that.
[00:02:22] And it was also, if you've ever played Pokemon, I played Pokemon as a kid. I really liked the nutrex of the idea of you just in the wild and you take a photo of a Pokemon and it tells you about it. So that's kind of what we built it off. We wanted to make a pokedex for food because of course, the idea of Nutrify, if it was an ideal case, was to get.
[00:02:46] Some of the market share of apps like MyFitnessPal earn money from that. But the real one that I sort of wanted to do and still pursuing is put it onto an iPad and take it to schools and educate kids at schools in a fun, interactive way about different foods. And so we're focused on Whole Foods, so we don't scan barcodes in the app.
[00:03:08] It's just purely a computer vision play of, Hey, is this food? Real food, like not in a package. And then it'll identify it from that. So it's mostly for food education focused on Whole Foods. There's a consumer app and a iPad app. The iPad app is a bit more focused for the younger generation to learn about foods in a fun, interactive way.
[00:03:29] Dr Genevieve Hayes: Okay, so you've mentioned two projects here. You've got your Nutrify project and your, machine learning zero to Mastery Academy course. What specific steps did you take to validate those ideas before investing in their development?
[00:03:45] Daniel Bourke: Well, the nutrify idea was actually when I first started machine learning and I never had the I guess backing to myself to sort of really pursue it. But then I kept it in the background as I was learning more and more. But then the zero to Mastery I'd left my machine learning engineer role and trying to start my own business.
[00:04:00] I published an article on medium of what I've learned in my machine learning job. And then Andre my business partner in Zero to Mastery reached out to me and said, Hey, I'm a web developer. I make web educational courses. A lot of my students are asking for machine learning courses because this is 2019.
[00:04:15] Machine learning was seeing a really big uptick in popularity and not as much as what it's seeing now, but it was quite significant back then. I. And I was like, sure. Yep. I like making educational materials. He showed me the results of like what he earned through web development. I'm like, okay, if I can replicate that in some way with machine learning and data science from what I've learned, then that's a viable income.
[00:04:42] And so I did that, built the course. He was very upfront, was like, Hey, I can't guarantee this, but this is where our goals are. And I was like, okay, I don't really have much at the moment. I'm living off savings, so I'm just gonna take the risk here. And luckily enough that turned out to work and it's still working.
[00:05:02] So I kept pursuing that and then it got to a stage where I had made three quite extensive courses on machine learning and deep learning data science and they are now in sort of maintenance mode , I update them once per year of making sure they have the latest libraries and fixing errors and whatnot.
[00:05:21] But they now sort of run on their own in a sense. And then I was like, okay, it's time for a real world project. 'cause I've spent two years now making educational materials and I'm a real big fan of the balance of like. Not just being a teacher, but working on the things you're actually teaching.
[00:05:36] And my brother is an iOS engineer, and so I'm like, Hey, do you want to just work part-time together on nutrify the app? And I'll build the machine learning models. You build the app interface and I get to practice now building a real life project.
[00:05:52] And what I've learned from machine learning, I can inject into the project and then what I learned from the real world of building that project I can put back into my teaching materials. And so it's a really nice symbiosis of having something in the real world that works, that people can interact with.
[00:06:07] And then taking the knowledge from that and then educating people, Hey, this is what it takes to build an actual real world model. And then serendipitously working on making. Computer vision models run on device really fast has led to another project that I'm partnered with a American company for this year as a consultant slash hands-on engineer for their computer vision project that they wanna run on a mobile device.
[00:06:33] And so the skills I learned from nutrify not only help teaching the courses, but I've also led to another opportunity of partnering up with a US company to build their computer vision models to run on device.
[00:06:44] Dr Genevieve Hayes: So what. I'm hearing is with the educational courses, your initial validation was people were demanding them, so you were meeting a unmet demand. So, that's clearly a very good signal if you're thinking about pursuing an idea and I. With Nutrify, it sounds like that was something that you were passionate about because of your background in food science, but it complimented your educational work so you weren't gonna lose anything from pursuing it, and it was gonna make that work more valuable.
[00:07:18] And that was effectively validated by the fact that even though this company you're now working for as a consultant, hasn't purchased Nutrify for a million dollars or whatever, or a billion dollars. They've purchased your services. So that's, indirect validation of the fact that you are on the right path there.
[00:07:38] Daniel Bourke: Yes. You're so right. So I probably went off a little bit on a tangent there to clarify that. Yeah. Nutrify, the validation for me was a personal thing of Hey, I want to build this thing. I've had this idea for a few years. I would hate to just let it go by the wayside. I want to just see it come to fruition.
[00:07:53] And you're so right. As a worst case scenario. It's been a educational process. I can take the learnings from that and teach them whether it's a win or a fail.
[00:08:02] In a best case scenario, the upside is uncapped of if it goes viral, people really like it. Well, then it's another boom. But really the main driver was, Hey, I just wanna work on this for fun, and then I'm gonna learn a bunch of it because I'm gonna make sure it's in the real world.
[00:08:17] It's not just gonna fall by the wayside. I'm gonna see it through to the end and then. Whatever comes from that, comes from that.
[00:08:25] Dr Genevieve Hayes: Have you had any project ideas? Years that have failed validation.
[00:08:29] Daniel Bourke: Ooh. I have a long list of ideas and notes of things that I'd like to work on. However, I would say since starting working for myself, I've kind of been lucky in the sense that the ones I've put a lot of effort in have. Off in a sense, and I'm kind of, I guess, biased here in the sense that I actually don't spend too much time validating aside from whether it's something that I'd really like to pursue. The validation, I guess for me was seeing the results already at the start of creating the course of like, Hey, if I can replicate half of that, or a quarter of that, then that's great results. It turned out to be, I guess a good break of like, okay, it worked quite well.
[00:09:15] But then other things I've got quite good at like saying no to or. Not taking on something that I really don't know that I can invest all of my energy in, if that makes sense. That's probably a critique of myself actually now that I'm thinking out loud about this maybe I should try more things that won't work.
[00:09:30] But again, I don't think many things don't work. So I'll write a novel, for example. A couple of years ago, and that was again, another thing for me. It wasn't selling like Harry Potter, but that wasn't my goal. My goal was just to write the book and improve my writing skills.
[00:09:46] And so to me, even though it wasn't sort of by a lot of people's metrics, a runaway best seller, to me I went from. A blank page to a published novel that someone can buy a hard copy of and read. So that was my validation for it. It was like, I'm just gonna see this project through to the end.
[00:10:03] So, I guess that's just my bias of I don't mind something not being super successful because I know that I'm gonna learn something from whatever project I decide to put enough energy in.
[00:10:16] Dr Genevieve Hayes: So what I. I'm hearing here is look for projects where there is evidence of someone having succeeded. At it previously so that you can say, well, if they can do it, I could potentially replicate that success and also look for projects that are win-wins where you either succeed and make lots of money or.
[00:10:36] Worst case scenario, it might not succeed as well as you want, but at least you can learn something which you can then capitalize on, or just look for things that are personally validating to you.
[00:10:47] Daniel Bourke: I think that is a perfect synopsis of it. It really depends on your goal. If it's financial, you need financial returns, well then sure, yes. Put a lot of validation into that. Or partner with someone who has proven success there. And then on the other side of things, create a win-win scenario where you can't lose of, like if you wanting to write a book, you get better at writing, you publish a book, you can say, Hey, here's what I've done, or you wanna work on your own project. It's like, okay, I'm gonna upgrade my technical skills. I'm gonna learn something new.
[00:11:16] The upside is uncapped. If it goes wildly successful, that's awesome. Financial return. If it completely fails, well it's not really a fail because , you've built it through, you've learned the skills to go through that. The next project, you can go, okay, I'm gonna approach this slightly different.
[00:11:29] I.
[00:11:30] Dr Genevieve Hayes: Yeah, it's like the advice I got when I first started this podcast was the secret to making a good podcast is to start by making a bad podcast and then make it better.
[00:11:39] Daniel Bourke: Exactly right. The first a hundred. It's almost like a numbers game at the start. You gotta just create the sets and repetition. Every gym goer kind of knows that. Intuitively the first day you go to the gym, it's not very good.
[00:11:50] But then a year in, it's sets and repetition. So creativity, projects, business. I'm very simplistic in that way. I think it's just sets and repetition. You want to get good at something, you just practice that.
[00:12:02] Dr Genevieve Hayes: So that's a wrap for today's value boost. But if you want more insights from Daniel, you are in luck. We've got a longer episode with Daniel where we explore Daniel's path from traditional data science employment to building a thriving independent career and discuss the practical steps needed to transform your technical skills into sustainable freedom and financial success.
[00:12:26] And it's packed with no nonsense advice for turning your data skills into serious clout, cash, and career freedom. You can find it now wherever you found this episode. Or at your favorite podcast platform. Thanks for joining me again today Daniel,
[00:12:42] Daniel Bourke: Thanks, Genevieve. It was a lot of fun.
[00:12:44] Dr Genevieve Hayes: And for those in the audience, thanks for listening.
[00:12:46] I'm Dr. Genevieve Hayes, and this has been Value-Driven Data Science.