Small Business Founder: AI Soccer Highlights App!

Transcription

(Music)

Dan: You have a really solid foundation when it comes to the AI stuff, get that. Talk to me about actually creating the app. Can anybody do this? Or is this, you need someone that actually has a coding background? Today on Time We Discuss, I wanna welcome John Poothokaran and it's time we discuss what it's like working in the AI space. John, thank you for joining me today.

John Poothokaran: Thanks for having me, Dan. It's nice to be here and I look forward to talking.

Dan: Absolutely, this is gonna be a good conversation. I find AI very fascinating. I like to say, it's a great tool that can do a lot of things that make our lives easier in a lot of ways. But when it comes to doing things kind of from the ground up, it might not be the best. But that being said, I wanna hear all about this project that you're working on involving, Americans will call it soccer, everywhere else calls it football. Talk to me about what you're doing.

John: I've been building Street for a little more than a year now. It's, we call the app Street and the company's name is streetplay.ai. And we're building AI systems that can help create a lot of football content for casual football or soccer. And what we're trying to do is, basically take a video of a football game or soccer game from any camera and then turn that into a match highlight, tell you the story of the game with the score and the progression of the game. And then also turn it into clips that players can share individually, as well as do player stats, game analytics and those kinds of things. All using an automated AI layer that kind of watches the video, understands the game and then gives you all of these different pieces altogether. So we've got an app right now on the Play Store and App Store and people can upload any casual game they play, just set up your phone on a tripod or have your friend record the whole game or most of it. And you can get your clips from the app. So yeah, that's in a nutshell what we're building.

Dan: That sounds really, really cool. And we were talking a little bit before I started recording and it seems like you don't have really any experience with actual programming. You did not go to university for programming, it sounds like. Talk to me about your education and how that kind of leads you into doing what you're actually doing.

John: I did my bachelor's degree in aerospace engineering at FIT in Florida Tech. So at the time I was mostly doing a lot of hardware, aerospace like design, aerospace systems, design and testing. And a lot of it was quite research focused like imagine doing an experiment in a lab. And that was kind of what kept me kind of busy and like the kind of different projects that I worked on during my undergrad. They were all building something. So built like rockets, planes, drones and like hardware systems for research labs. So through all of that, it was definitely maybe the time where I wasn't programming as much, but still got exposed to computer science and programming through like C,++ MATLAB, all the basic stuff that engineers kind of use without going too deep into CS. So yeah, after that, I went to Stanford to do my master's in aerospace engineering. So I wanted to kind of continue the story, see where it would lead. And that's as I was kind of looking into different spaces to work in, I took a few random classes. And one of that was an AI course called Decision Making Under Uncertainty. And it basically introduced me to like patient systems and reinforcement learning. And also a lot of different programming like all at once. And I got interested in that. And as I was exploring what I wanted to do next, there was not as much programming experience, but a lot of the math kind of translates over from the stuff in engineering. So it was a bit of a difficult step, but I had the foundations that I needed. So then during my second year at Stanford, I took all the AI courses that people were offering, like the courses that had like the lectures halls full and kind of overflowing. I would like kind of sit in at least the first week or two and then pick my favorites based on what I saw. And that's when I got into the broader like machine learning and AI space. And at the time, like NLP, which is like large language models, you know, natural language processing, that particular field and computer vision were kind of taking off. This was around like 2017, 18. And in 2012 is when computer vision really started taking off with neural nets. So there was a lot of exciting stuff going on in neural networks and kind of the research and the product side of things. So I kind of got interested. And what really hooked me was how word embeddings work. So there was this lecture that I sat in where I-- it was just introducing me to word embeddings and how semantic meaning was captured and just a list of numbers. And that was really interesting and kind of unusual. And that was kind of what got me hooked into deep learning. So yeah, after that, I worked as an AI researcher at a cybersecurity company at Avast. Now it's called GenDizzle. And yeah, we worked on kind of NLP focused product research for like cybersecurity, privacy, and insurance. So that was five years at the company. And I got to work on a lot of new different projects that basically took the deep learning technology and latest advancements and then applied them to specific problems.

Dan: OK, so you have a really solid foundation when it comes to the AI stuff. Awesome, get that. Talk to me about actually creating the app. Or do you have a partner that's working with you that's doing the actual coding? I'm very curious about-- can anybody do this? Or is this-- no, you need someone that actually has that coding background. Talk to me about that.

John: I think with chat GPT and a lot of the coding tools now, it becomes a lot easier, especially if you have some experience in a particular programming language. It makes you much, much faster. Because now you're just kind of-- I would say even from last year, the tools are much better now. Now I just need to read the code and kind of accept the changes. But yeah, the main part is kind of understanding the customer. I started working on the back end code last year. And my co-founder and CTO joined like three months ago. So I was working and built kind of the MVP working version without the app, but just the code that runs the back end. And then got help from my brother, who's a software engineer. And he built the app. And my co-founder and I both work on the back end and AI systems. So yeah, I do have a co-founder. It's super helpful to have a partner. It's difficult enough as it is to build something from scratch. And there's a lot of failure and rejection involved. So it's super helpful to have a partner to go through all of that. But yeah, I think with the tools these days, especially to get a prototype or something running, it's very-- yeah, it's basically how much can you sit down and talk to the AI about this? Because eventually you will be able to build, at least the first version of what you want. And then obviously you have to keep learning from there. But definitely I think it's much easier than before to start.

Dan: Talk to me about football. Why did you-- I'm assuming it's probably a passion of yours, a hobby of yours. But why football? Why did you decide to go down that road?

John: I've been playing football, soccer since high school, and all of it casually. The most competitive leagues I've played in are like intramural leagues and just playing with my friends on a team. So yeah, it's been a lifelong thing. It's survived many career changes and changes in my life. But I've been playing football at least once a week for-- yeah, since high school, through college, through life after that as I worked and stuff. And I still do. And about sometime in the summer last year, I really wanted to build a few products. Basically because after chat GPT happened, I feel like it took away a lot of the excuses that I might have. Because now you can build things much faster. So I definitely wanted to try a few things. So I did try a bunch of ideas that extended from my previous work. So I tried to work on misinformation using LLMs. And then a couple of other ideas that I was prototyping, like an education website where you could go and basically get guided by AI through many different topics, but then also help you create your own curriculums. But I saw a lot of people who were already doing that. And then I had this idea of where's the football app? Where's the soccer app? Runners have Strava, but where's the quintessential soccer app that everybody talks about that keeps the culture going? That's-- there is no soccer app. So I thought, what would that look like? And the thing about soccer is everybody plays, but in a 60-minute game, you only touch the ball about two minutes if you're playing. And even at the professional level, even Messi is touching the ball three minutes out of 90. So it's really a very social game. It's really about being there with your teammates. But it's also about, I feel like a lot of people who play soccer, it's all about the moments. Like the game, there can be a nil, nil, zero, zero game. But the excitement can happen in just 20 seconds, 30 seconds. The game can change. So I think soccer is a lot about moments. So I think that's the first thing I was thinking about. The app would be about the moments in football. So yeah, and that's kind of where our first product, which is highlights, which it's a bit unusual, because a lot of sports companies, especially companies that are trying to get into this space, a lot of them do analytics. But we're very focused on just the moments part of it, making the content really nice, making it really easy to get the content out of a game, and kind of taking away that blocker of the difficulties of recording and editing, and just making it that much easier to share. So I feel like a lot of people leave the soccer player that they are on the field. But now you can take it with you and share. So that's kind of where the idea comes from, what I'm trying to build.

Dan: If someone wants to do something similar in general idea, on a macro kind of level, OK, I want you to think about the roads you went down. And I want you to think about-- because when someone's starting a business, you do all kinds of research. You do your market research. You check to see if the product is viable, this, that, and the other. What was one of those barriers that you kind of came across where you weren't expecting it, where you felt like you were prepared, and there was just something that kind of like, oh, I hadn't thought about that, that kind of caught you off guard. What would something like that be that other people could learn from?

John: I think the most important thing is kind of the team you work with, and having people that you can rely on to kind of help you through the journey, especially in the early days. So I think really kind of getting your support systems in order is really good. For example, just having friends that are experienced or have some expertise that you're lacking is super helpful. And just kind of reaching out, getting an opinion even can be very helpful. Like I've had calls with friends who have previously started companies or are currently doing so. And just from that one call, it would be clear that I'd have to make a pretty serious decision or at least change something that I'm doing. And I think that type of the change in your action, you can't always expect it from you. If you're the one with all the ideas, then you're just wasting a lot of ideas that other people might have. So I'd say definitely get people around you, and that can really help. I think that's a good way to kind of keep going even when things are a bit tough. I think also with a new product, obviously it doesn't come out like the polished final version. So there's a lot of people that just kind of naturally don't want to try a new product. Like a lot of people are very comfortable with products that they've heard of many times, but something new comes up and they're not really open to trying it. So it's very natural for it not to go well. So I think, yeah, remembering that this is just the very, very beginning of something very specific that I want to build just kind of keeps me kind of knowing that there's more to come, I think. I think that's the kind of headspace I try to be in when a particular thing doesn't work out because you've got to keep going.

Dan: I love so many things that you're saying right now. The one thing that made me think of-- so you talked about reaching out to friends and kind of leveraging their expertise. And that made me think of this question. It sounds like that you didn't go to school formally for any kind of business degree or anything like that. So the one question I'd like to ask when I'm speaking with an entrepreneur, what did you do to kind of fill in those knowledge gaps when it comes to starting the business, when it comes to marketing or website design or something like that? How did you level out the business side of things?

John: These side of things are still not leveled out. But I would say definitely I've been kind of self-studying for a long time. This is not the first project or something that I worked on that I wanted to be a startup. At Stanford in my second year, I worked on an edtech project that I worked on it for a quarter. We got into one of the startup classes, Lean Launchpad, but I wasn't able to join that course at that time. And since then, I've been trying. I've tried an agricultural crop monitoring startup for a year during the pandemic. And then for another six months, I was also considering something in the blockchain crypto space. And all of these things were kind of just projects. It was like either during school or during free time outside of work. So they were all kind of just basically introducing me to the world of customer and user interviews and kind of that process of just trying to do sales or business development. I also did an internship at StartX, which is a Stanford-affiliated startup accelerator. And my internship was basically B2B sales, looking for partnerships for the accelerator. So I had some experience that I would piece together, like sales. And then from work, I really learned a lot about product just by interacting with more product-focused people. And I would read a lot of product management books and MBA-related books, because I did plan to do an MBA, but I realized an MBA is mostly about networking. So you should kind of go in with a lot of the knowledge that you can read beforehand. So yeah, I've had a lot of time outside of work and outside of my main stuff, just kind of brushing up on different things. And yeah, and then now I think with design and website and app building, I still think there's a lot of space for expertise. So I definitely reach out to people where I need their help. But yeah, a lot of product and sales work. I've kind of at least tried to train myself up even before this. So I feel comfortable because I think I enjoy it. I kind of know, I at least have the previous experience to compare it to, like, I know it's going better than the last time.

Dan: Let's talk about a typical day for you. Are you, when it comes to this, to your company here, are you marketing a lot during the day? Are you on social media a lot, trying to promote it during the day? How do you spend your day right now at the level that you're at?

John: Right now it's a super small team. So it's just me and my co-founder full-time. And then a few people that we work with kind of part-time or like, or on a contract basis. So it's very much, we're running very lean, where it forces us like every month, kind of the number of games we serve increases. So it forces us to increase our automations that we have in the workflow and kind of reduce the effort you'd need from one person. So there's a lot of that, where earlier it was kind of talking to a lot of people and developing the code base. Now our highlights and content product is kind of serving a lot of games and it's growing slowly. So as we do that, we're trying to speed it up, improve it, reduce the cost, make it more automated. So I do a lot of work that reduces my future work on that front. And I work with my co-founder on figuring out, kind of where we are with like, the data is really important, we're curating and building our datasets. So kind of planning that out, because once you set a plan, like the dataset only appears like, maybe like a month or three months in the future. So there's a lot of kind of figuring that out at the point that we are, where we're trying to do a few new things. And then yeah, a majority of my time is definitely doing sales, which is either outreach on Instagram or LinkedIn. And then also meeting people in the ecosystem over here. So yeah, that's kind of how I spend most of my time. I think now that we're interested in fundraising this year at some point, so I'm also trying to reach out to investors and build relationships that way.

Dan: Earlier you talked about different projects that you're working on. You talked about, there was the blockchain one, there was another one you said during COVID, I think you said, how did you decide that Street was able to go to the next level, that this was business worthy? Was it because you were more passionate about it? The other ones were legit more projects, they were just a means to get you somewhere else. They didn't have legs to stand on. What made this one different from going to, from project to business?

John: I definitely kind of took the leap before deciding that, like I knew I wanted to work on this concept and this product, but I took the leap kind of giving myself a year to figure out what I was doing. And part of that year was just kind of having very low expectations of leg success, but very high expectations of exploration. So the main goal of the year was just to understand if this thing had legs. So I did spend a lot of time kind of speaking to a lot of people, playing soccer myself, and kind of putting myself through the experience of getting my own highlights and the people around me. And there was a lot of that time where I was just, yeah, basically talking to people, understanding the experience and seeing if it had legs. And then, yeah, I think I was, building more conviction just in the experience and then somebody wanted to pay me for the first time. And that was when I realized, okay, somebody wants to pay. Now there's something here at least. It's not just a free app. So yeah, that was the first kind of inkling I had. And the other thing is, this time I had decided I was just going to quit my job and kind of devote full time to building a startup. And so I had a lot more time to put in the reps. The math I always used to do in my head when I was doing the other projects was, if you're working five days a week, and let's say you take another half day for all your, just maintaining your life, and you have a day and a half really that you can put into this thing, that you're like your project. And if you do that over a year, you get about 60 days or like 75 days, which is less than three months. So the math, that just means you're never going to make that much progress. Like compared to if you just quit your job, which is a very tough decision to make, but I really wanted to do it this time. So I did. And I think that progress that I never saw in the other stuff came through because I had that much more time working on it. So I think there's a huge difference there just in terms of commitment. Obviously the experimentation requires some time and that really helped. And now, yeah, like I've got a real understanding of the ecosystem, maybe not a full understanding, but much better than last year. So I understand like a few of the customer types that are involved and exactly what they've been asking over multiple conversations. So I can see that there's, I can see reasons to be working on this and reasons that the way I see it happening can happen. So yeah, I think, yeah, you know, all the other projects combined probably I worked on maybe like three different projects for like maybe three months of just work time. But this one, the first time it's gone like more than a year, like full-time. So I think that's where a lot of the progress is. It's after a lot of the bad parts.

Dan: If you had the opportunity to give some advice to yourself from the past, so yourself a year ago, what would that advice be?

John: Keep doing a lot more sales. Always be reaching out. I think in the beginning, definitely there was some stuff I had to build to kind of have a conversation with people and offer something of value, but definitely more conversations earlier because a lot of the information is just not with you in the beginning. So you gotta be going and collecting information. So yeah, just yeah, a lot of that, take a lot of action to get information. And that usually means talking to people, whether it's people who might buy from you or just people in the ecosystem, or just other people who are like the audience of people who watch football, for example.

Dan: John, I feel like we could easily continue talking about this, but unfortunately I am running out of time. One thing I like to do whenever I'm speaking with someone, I like to give them the opportunity to talk about a project they're working on, the cause they believe in, more about their business. You have quite a bit to choose from, but there's something specific you'd like to discuss. Floor is yours.

John: Obviously I've been talking about Street. So if anybody plays football, soccer, or futsal, any form of the game, and you've got a chance to record your game, try it out. Try Street, try to get your highlights. Yeah, otherwise a cause I really like is I really like Wikipedia, because at least, I don't know if it's as relevant now because you can ask AI stuff. But Wikipedia was the website that in high school and college would really kind of help me explore my curiosities and a lot of science and engineering related fields. So yeah, if you guys like Wikipedia, check it out. Donate if you can, that's nice. But yeah, Wikipedia is super nice for curious people.

Dan: John, that is awesome. I'll see if I can get a donate link button for Wikipedia. I'll put it on the website. Also get links for your business, your app. I'll put them in the description on the website. They'll be all over the place where people are trying to find you. They can easily find you. It's been absolutely awesome having you on time. We discussed and we learned what it's like starting a startup.

John: Thanks so much Dan. It was super fun talking.