Jake Aaron Villarreal: Welcome to our podcast, Born in Silicon Valley, where we interview startup founders exploring their journeys, their success, challenges, and lessons learned. We hope you be inspired discovering what it takes to build a thriving startup. I'm your host, Jake Aaron Villarreal, and here with us today, we have Jeff Lu, who is the founder and CEO of Akool. Jeff, welcome to the show.
Jeff (Jiajun) Lu: Yeah, thank you for inviting to the show.
Jake Aaron Villarreal: Yeah, well, thanks for jumping on here. Uh more about Jeff. He's not only the founder and the CEO of Akool, which is a leading generative AI powered personalized visual marketing platform, but he's got a blend of background in technology and business, and has spent the last decade working on making videos and images look better with AI and has obtained a PLDA from Harvard Business School and has a series of successful startup experiences under his belt. So Jeff, uh, we're gonna talk about your company and kind of what you're building, the problem it solves. But before we do that, where are you calling in from today?
Jeff (Jiajun) Lu: Sure. Yeah, I'm calling in from, uh, Santa Clara, California. Yeah.
Jake Aaron Villarreal: Ah, very cool. I know that town. I know that area. Silicon Valley. It's heart of technology. Are you originally from there?
Jeff (Jiajun) Lu: Not really, but I moved here for seven years and before I was in Illinois. Yeah.
Jake Aaron Villarreal: Gotcha. In Illinois before that. Okay. Very cool. All right. Well, um, you know, a lot of founders we've had on the show kind of have a DNA of startup experience in their belt even as a young child. Uh, be a lemonade stand. It could be selling services or products to their friends in school. What was you, what would, give us a little bit of a shape of your background in terms of where your entrepreneurial spirit comes from?
Jeff (Jiajun) Lu: Yeah. So when I was young, I want to be a scientist in physics. But uh things change when I read a book in high school. It's a book uh mainly about all the different companies, how they grow and how they expand. And uh I feel really amazed about everything happening in the business world. And that's the first thing that got me interested and want to, want to do something myself. Yeah.
Jake Aaron Villarreal: Well, and what was the name of that book? Do you recall? I know it's a long time ago probably, but anything that would be worth listening to or reading.
Jeff (Jiajun) Lu: Yeah. So, it's a book from um China. So, it's mainly about the business world in China. It's the book name is uh Big Failure. It describes um like 10 to 30 stories of the largest companies in China, how they, how they grow up and then how they, how they crash. But it's a, it's a very interesting story. Yeah.
Jake Aaron Villarreal: Yeah. That's great. Yeah. Inspiration could come from anywhere. It could come from a book, your parents, your friends, an experience you have. Uh I'm a huge believer in reading and listening to other people's experiences, biographies. If you can fast-track how you build business by someone else's failures or experiences, uh you know, what better experience can you get from learning rather than going through it yourself? You know, this podcast is about trying to, trying to reduce the time it takes to build a company and learn from others that have done it prior to you. So, I know you have other experiences in the startup space. Give us a little background of your startup experience.
Jeff (Jiajun) Lu: Sure. Yeah. So, um since then, since reading the book and so on, I'm very interested in doing startups and so on. And actually, the region I'm coming from is also very uh, there are lots of businessmen. So yeah, so um the first company that I participate and uh uh is about nine years ago. And at that time we uh, two of my friends found me and said let's do a uh company for dating. So it's a dating app and uh it gather angel investment from one of the uh famous angel investors. And uh then uh I worked on that for uh less than a year. And um it didn't go that well actually, so I left the company. But that company actually now is doing very well. So it shifted gear and moved into the space of uh cameras, um like camera for blue collar workers. And it became the third largest uh working platform in China. Uh 300 million users and...
Jake Aaron Villarreal: Oh wow. Yeah.
Jeff (Jiajun) Lu: And 300 million users.
Jake Aaron Villarreal: Yeah. 300 million users. So that's the population of United States almost. So yeah. Um and the second experience I have is I participated uh a French company. It's called XREAL and it's a AR glass company. It's I think in terms of the number of units sold they are probably the largest AI class company in the world. Um and uh yeah, help them to do their AI ML in uh with their AI, uh AR glasses. Yeah.
Jake Aaron Villarreal: Wow. That's great. There's a big trend around glasses with you know Vision Pro coming out and you know what that could look like, how it works, operates. Is it for the consumer? Is it, can it be applied for business users? There's a lot in there. It seems like you're ahead of the curve on a lot of these technologies where you were working in maybe the earlier stages of the timelines of when these categories of products come to market. Uh specifically AI, it seems like AI is much more prevalent today, but you were working with it very much beyond the last 24 months, it sounds like. So, we're going to get into your background in your current company here, but um what inspired you to even kind of get into machine learning or AI, just in general the technology sector that you focused in on?
Jeff (Jiajun) Lu: Yeah. So, it's purely out of interest. So, because the field of working on videos and images is so interesting, a lot of things are very intuitive. So uh and I like physics at the very beginning and uh which means uh physics and math right, there are lots of math and so on. So I get into the computer science and the field that I initially got into is the most physics and the mass heavy which is computer graphics and also uh do physics simulations and all the mass formulas and so on. That's the, that's probably the most mass heavy field. And um in that field I start with 3D, 3D generation, how to generate 3D geometries of uh different objects such as chairs, airplanes and so on. Now like you just pass it through neural networks and one shot right, but, but like more than a decade ago it's not. So you, you need to do lots of mass modeling and also machine learning modeling as well for the 3D models and uh make them uh interesting. But later uh 3D and video images, they [are] so close, and a lot of the 3Ds are being used to create a movies and videos. So I, I work with my PhD from uh University of Illinois at Urbana-Champaign with a big name professor in the field and moved fully into the field of video and image creation. And all of these move and choose because of interest, because this field is very interesting. Yeah.
Jake Aaron Villarreal: Yeah. Uh there's a lot going on there. I mean, if you think about some of the biggest social networks, it's really about photos and images and, you know, sharing that information and, you know, engaging others. So, I know there's lots of opportunities in that space. Uh, talk a little bit about the company you have today, Akool, and what, what really was the inspiration behind it?
Jeff (Jiajun) Lu: Yeah. So, Akool is um the blend of their experience in technology, in business, and also the marketing mean. So, it's a blend of three things. Um for the uh for this field, I've been working for a very long time. So definitely it's a field that I, I like a lot. And the marketing is also very interesting, which uh I got from uh HBS. And I, I really like marketing a lot and currently also involved the uh marketing professor from HBS in our company. And uh the last thing is market need. So we initially uh more like uh more general platform and uh uh go to the market with the technologies with the experience that we have. And then we really get a huge amount of traction from the marketing perspective. And we, we shifted everything into marketing and go deeper with the marketing use case. And I think in United States marketing is also the largest uh largest area for IT companies. Most of the large companies and so on giants, their majority revenue source is marketing. So yeah.
Jake Aaron Villarreal: So if you're a creator or a marketer out there and you look at your, your, your company or your platform, what do you give them? What's the value that they get if they were to use your platform?
Jeff (Jiajun) Lu: Yeah. So the problem we are solving is marketing is so boring, right? People, to, people don't like ads anymore because uh ads becomes more and more and nobody wants to watch ads. And the value, the things that we bring from it really make the marketing more interesting and more engaging to make uh people willing to get engaged, willing to watch ads and willing to do uh to do things with it. So um I think that's the biggest value we bring forward for, for this uh, for this field. And uh how to do it is we mainly help them to create uh interesting advertisements and interesting ads and very personalized ones, which means uh you see yourself in the ads, or you might see the person in the ads talking to you. And uh you, you can, you can see uh person the ads that to your taste and so on. So, it's very personalized uh advertisement and marketing to get people engaged. And it's also very useful for creators as well. There are a lot of influencers that they want to create engaging content um to make people interested. Uh and even for localization as well. Like you want to uh speak different language, for example this podcast it's in English right, but if you want another podcast you want a Spanish version right, so, or, or Japanese version. So you can create uh different versions of the podcast and get, get more audience. So I think it's definitely a very exciting field and lots of fun things are happening. Yeah.
Jake Aaron Villarreal: Yeah. I mean I'm on your website as you're talking and I, I, I, I love what I see. I mean for example there's a picture of a face and you can swap, swap that face. So you can go from a white person to a Latino to a Asian and very quickly it looks like. Or you can take a picture you've already, maybe it's a product uh you're a makeup product and you want to change the background very quickly. And looks like you could do that uh seeing avatar that you're basically typing in what you want that spokesperson to say on a video which looks like a very realistic person. And I'm assuming that you just type in what you want to, want that person to say and it's going to say that message. So, if you're doing an ad or something, it would be like you have a spokesperson. I'm kind of talking through this kind of, I want the listeners to understand what you actually get with your platform. Uh, talking photos. I'm seeing the translation of like a beautiful woman. It looks like she's a spokesperson going to talk about something. And I'm assuming that with a video translate you can have her speak in multiple languages. So, this all looks really, really cool. Uh what's the time it takes to do the transitions from one language to another or from one phase to another? Like how fast is the technology and is it all sort of AI built out systems that you've created over the last several years?
Jeff (Jiajun) Lu: Uh yes, that's a very good question. So we created the technologies ourself. Um which means the architectures, the data we collected and all the trainings and so on. And the goals that we do is really about to ensure the quality. So uh we optimize for the studio quality. So lots of our data is 4K and so on. And also we ensure the um ensure the eye-catching part because for stu- uh for, for marketing advertisement is very important for your videos and images to be eye-catching, which is lots of the studio settings and so on. So that makes it very different. And we do the copilot ourselves from uh data uh collection and the labeling and the training and so on, which makes us pretty different from, from others on the, on the market. And the whole process is usually uh from three months to six months to do the, to do all these works. And the processing time, because we use the largest models and so on, the processing time varies. Frequently you see one video, one minute of video will be processed for three to 10 minutes. The processing time takes uh takes long uh relatively longer because we use larger models and uh so on. So, and also processing higher, with higher qualities, with higher resolutions and so yeah.
Jake Aaron Villarreal: Yeah, I'm not doing it justice by telling you about it. So for the viewers, the listeners, you know, go check out Akool, akool.com. And when you go there you can see it's beautiful design but you know, 150 million assets created, you know, companies like Coca-Cola, Amazon, Microsoft, Autodesk use the platform. So, it's, it's not a, you know, fly by night, you know, platform that came up with AI in the last couple months. It's something that's been around for a long time. Obviously, it's got a proven business model and it's functional and it looks fantastic when you're just looking at what you've created. You're obviously ahead of the curve when it comes to this type of opportunity to get into the marketing side of things here. What's been um some of your sort of bigger successes that you could talk about of companies that have used your, your platform? And uh what type of ads or what type of, where are they putting the ads once they create it with your platform?
Jeff (Jiajun) Lu: Yeah. Yeah. So, we worked with uh many top brands and top celebrities and much more to come. Uh one of the interesting things that we did uh that really go viral at amazing speed is uh we work with uh a, a cookie brand from India. It's called Sunfeast. And their um their brand ambassador in SRK. So uh the Tom Cruise in India. So um so he, he, he is over there. And so what we do is we launched an advertisement campaign. So everyone, no matter males or females and so on, can be in the commercials together with SRK. And um they can generate the videos, share on the social media and so on. And that thing really go viral at an amazing speed. So they launched the uh, after three days they launched the campaign, the, the usage goes up to 20 people per minute. And like after seven days the usage goes up to almost 50 people per minute. That's, that's crazy number. So uh you think the YouTube video upload is 200 per minute, how many people upload videos to YouTube? So it's, it's crazy number. Um and end up they do traffic control. They cannot hold the traffic. And it happens so fast. It only take, take one week and the traffic exploded. So yeah, that, that's something that we believe it's, it's very interesting and that it goes, it goes viral very fast. Yeah.
Jake Aaron Villarreal: So if you want to create stuff that goes viral, that's, this is one way to do it. Uh that situation, that, that, that use case you're talking about, uh so is there a level of buy off that, that famous person needs to say, "yeah, you can use my persona for this ad campaign." And then are they part of that campaign too, or is it just "use my clones"?
Jeff (Jiajun) Lu: That one is uh, celebrity is originally together with the brand. So the celebrity is original partner with the brand and they shoot additional materials. I think some, I think there are at least two videos. It seems one video is, is an ads for a while and they just you reused an ad. Another video is they shoot a new one and they use two videos for males and females. And so the celebrity is original in the game. So originally they win the brand. So yeah.
Jake Aaron Villarreal: Give me a little bit of um your perspective of there's you know other technologies that have come into this space. And you see obviously a lot of different platforms in marketing. What makes Akool very different, or I don't know, in some ways maybe more specialized than other companies in the space.
Jeff (Jiajun) Lu: Yeah. Yeah. So we, we are pretty special I think uh in terms of uh what we do and in terms of our uh team and technology and so on. So I think currently in the marketing advertisement space uh video especially personalized video is uh is not a thing yet. Um and we are like leading the frontier of it. So before there are some startups trying to do text and image, which is significantly easier and significantly cheaper to do, but we are the almost the I think we are the first one that's um doing videos in their marketing advertisement and do it at large scale, which makes us um very special. And also in terms of the technologies we develop, we develop the technology from scratch by ourself. Uh we know there are a lot of open source things out there. And if you track, a lot of people use open things uh solutions and just launch and provide it for free or very cheap. But we are not that player. We, we, yeah so we develop the technology ourselves uh and ensure the premium quality, and we charge. So that's our business model. So that's why we go to the more uh commercial and business side because most of the random users their will to pay is relatively low compared to the business uh users.
And also our team is very experienced. So we have uh great researchers, engineers and also business person uh and uh advisors on board help us to uh to navigate the company uh through the, through the market. So uh lot of the fun things and interesting things happening. Uh and definitely we believe that we are pretty unique on the market and we are doing pretty well as well. Yeah.
Jake Aaron Villarreal: As a company, are you bootstrapped or did you get funding? Have you gone through an accelerator?
Jeff (Jiajun) Lu: Yeah, so the company is almost bootstrapped. So we get some, we get some investment from some people we know. Uh we, we didn't raise too much money. So the company is, the first year of the company we can show the burn rate to be very low, and from the second year we begin to be uh break even and profitable and so on. So uh yeah.
Jake Aaron Villarreal: It's really cool. It's the way most companies are being looked to build their companies today. You know, path to profitability is line one now. It's not just "get the funding and go build and find your market." It's "find your market, do your research, make sure it's a viable product, get the funding, you know, and grow and scale but profitably." So, uh, really cool as a company. In terms of accelerators, that did you go through one? Are you going through one?
Jeff (Jiajun) Lu: Oh, actually we are in several accelerators, but one of them are like not so official accelerators. So, we are in several together. We are in a brand which is uh and Daniel Nat is the former GitHub CEO. And we are in Nvidia's accelerator. We are in... yeah, we are in Amazon's accelerator, and we are in NYU's accelerator. So we are in at least three accelerators. Maybe there are more, but most of these accelerators don't have too much things, but we are in the accelerators.
Jake Aaron Villarreal: That's great. Yeah. Hey, it's, you know, it's, it doesn't hurt to get advice from others as you're going, going through and scaling your company. Um, you taking that information and, and hopefully being able to contribute back at some capacity. Um, what's your trend uh your view on the trend of AI today with generative AI and you know everything that's come out from OpenAI to ChatGPT, and where do you see things going from your perspective?
Jeff (Jiajun) Lu: Yeah, so I think it's just the beginning of the generative AI. So this field is uh moving really fast. So it's like a, a decade journey. And this um, so the, this is a second wave of AI right for, for recent uh few years. The first wave is probably about 10 years ago, and this is the second wave, and this wave will be significantly bigger than the first wave because it, it touches almost everything that uh people do. So uh in terms of ours, we also think it's the beginning of our journey as well because uh marketing and advertisement uh is, the personalized worms the market, so, so big. And I think the current market penetration is less than 1%. So majority people are still not using it yet, and it definitely will change a lot.
And the, the, the move of the generative AI really depends on the uh cost of the computation and the availability of the data, and also the breakthroughs in the model. The usual model breakthrough I think each term probably each new breakthrough, a big breakthrough is like a decade right, one 10 years. And uh, but the breakthrough in uh breakthrough in the computation is significantly faster, and the breakthrough in the data is also significantly faster. So with cheaper computation and more data, the model could be better and uh things could be applied at scale. So for example, if you are able to create personalized videos at scale, real-time, at a very cheap cost, then I think almost everything you consumed online will be different. Everything will be personalized. Even every YouTube video you watched will be in your language and will be uh with more aspects related to you. A lot of time generate, a lot of things generating in real-time. So all these things heavily depends on the computation and the data. And the computation and data is um growing really fast. And the technology usually grows in different states like in terms of every 10 years or something, but computation and the data is growing really fast. So we, we do believe uh the penetration is still very low, less than 1%, and for the next 10 years it will uh grow very fast and it will rip the landscape. Uh now most of the internet companies uh are content companies. And when AI is the doing content, totally different, there, there is a lot of opportunities for startups and so on. Yeah.
Jake Aaron Villarreal: Talk about your company a little bit. Where, where are you headquartered? How big are you? And are you an on-site company or distri- distributed?
Jeff (Jiajun) Lu: Yeah. Yeah. So, uh company was formed two years ago, uh Silicon Valley. So, uh our headquarter used to be in downtown San Jose, just moved to Sunnyvale. And uh have about 50 people. Uh is about 30 and part time about 20. So about 50 people in total and it's pretty globally distributed. So uh most of our senior person are in uh in the valley. So but we have people globally from different regions and places. So yeah.
Jake Aaron Villarreal: What's worked for you in terms of hiring remotely people that are in different countries?
Jeff (Jiajun) Lu: Yeah. So uh there are a few reasons that why we hire remotely. First, the COVID totally uh [changed] the style of work right. So even, even in the, in the valley now, still a lot of companies goes two, two days or three days to office every week. Um and uh there are also lots of the permanent remote things. And for us, after the COVID we feel like uh remote work is a pretty normal thing. And the second thing is about the cost as well right? So uh in the, in the, in the Silicon Valley, the cost is pretty high to hire uh hire persons. And even these uh big tech companies are moving to Texas, Austin and so, so on. Um so we uh definitely leverage uh the um the uh better uh better cost structure by using um grow resource and remote resource. Yeah.
Jake Aaron Villarreal: Yeah. Really cool. How do you manage the performance of the team? Like how do you know when someone's really performing or maybe they're, they're not contributing at the level you know they should? What type of tools or infrastructure do you have in place to make sure that you're keeping your employees operating at a, at, at a maximum potential?
Jeff (Jiajun) Lu: Yeah, that's a very good question, and we also have lots of discussions around that because the team grows uh really quickly. So at the beginning of 2023 we are like a team with 15 pe- 15 people. Uh we are team with around 50 people. So when it's a team of 15 people, like I talk with everyone and get in touch with everyone and so on. But uh now we, we have a middle layer manager and I don't get chance to talk to everyone so frequent. Um uh we use, um we use some tools to manage the team like uh Looker and also uh suites. And uh that's the main thing that we use to do the team management.
Uh still a lot of things we are still figuring out. And by my, my personality also very hands-on. So I actually, actually try to understand everyone's work, and for some of the work I even try to do it myself. And uh better understanding of what's happening and so on. So a lot of the, because my background uh blend of technology and business right, so uh only thing I'm not that familiar with is front-end development, and besides that I almost do everything else. So from algorithm to backend development, uh PM works, some design works, and uh marketing works, sales works. I almost touched everything except for front-end development. I have no idea how to do that. So uh uh the work even I will do it myself, especially if the t- if the person we hire not too senior, I will do it myself first and then I figure out how to do everything and then I just uh teach the person to do the, to do it. So, so which means it's a, it's a pretty interesting thing and I'm, I'm pretty uh get in touch with everyone on the team. Yeah.
Jake Aaron Villarreal: Yeah. That's great. Yeah. There's always different technologies and processes to understand how you could manage your team that's remote, which is a different world today. How do you do it differently, better, faster, and keep people accountable? Uh actually went out and did some research as well and found some really cool companies that are doing this well, where they can manage who's doing what, how often, why are they performing better than other employees. And you could really kind of track what people are doing in a good way, but also coach them and be able to implement ways that they can be, you know, upleveled from a, you know, a C player to a B player or B to A. So really cool technologies that we've seen and we'll have on the show as well at some point. But u- terms of 2024, what are you excited about? What do you see on the road map for your company?
Jeff (Jiajun) Lu: Yeah. So we have a huge road map. So the, the thing is about execution now. So I really excited about the market, and we really believe the um, lot of things are happening. Um so uh basically for ours, in this year we want to, we wants to grow really fast. Like what we did in 2023, and we want to provide a better uh product to all the users. Um and uh still we always did the best quality results in what we have been delivered, and we will keep delivering that. Yeah.
Jake Aaron Villarreal: Yeah. That's awesome. If anybody wanted to find you, Jeff, or find Akool, where do they go?
Jeff (Jiajun) Lu: So, they can go uh akool.com to find us. And if they are interested to talk to me, they can write me an email at lu@akool.com. And akool.com.
Jake Aaron Villarreal: Uh based in Silicon Valley. Jeff, excited to see what happens in 2024. Be great to have you back on just to give us an update of the growth and uh the innovation that you're doing in that space, which I think every company should be looking at when it comes to marketing and ads and trying to take your brand viral, whatever your product is or whatever your service does. Excited to have you on with us today. Um thanks for coming on, sharing your story, and to all of our listeners for listening. It means a lot to me that you spent your time with us today. I'm your host Jake Aaron Villarreal signing off for now, but can't wait to catch up with you all on the next episode. Until then, take care.
Before we wrap up, I want to give a big shout out to all the entrepreneurs that have joined to make this podcast possible, and for all the listeners for listening. It means the world to me that you chose to spend your time with us today. I'm your host Jake Aaron Villarreal signing off for now. We can't wait to connect with you all soon on the next episode. Take care.
This show is sponsored by Match Relevant, a company that helps venture-backed startups find the best people in the market. And they do it in three simple steps. First, they sit down with founders to understand their story. Second, they tell their story into multiple candidate channels. And third, they schedule interviews within 48 hours. Find us at matchrelevant.com to learn more about how we do it.
.jpg)
.webp)
.webp)
.webp)
.webp)
