Lena Skliarova-Mordvinova, Good Face Project CTO and Co-Founder: Revolutionizing Personal Care with AI and Resilience in Challenging Times

Speaker A: From 7ctos. My name is Etienne de Bruyne and you're listening to the CTO podcast. Every week I spend time with fascinating people that enrich the lives of chief technology officers around the world. From perfecting the basics of building technology organizations to inspiring our minds into shaping our future. As always, the CTO podcast is brought to you by seven CTOs helping CTOs become world class leaders. Let's go. Okay, today in the CTO studio I have Lena Sklarova Mordvinova. She is from Ukraine. She worked at the Ukrainian Space Research Institute. Deep, deep background in AI and data science. But you might be disappointed with this conversation because we stayed high level in parenting, brain blending, America's trying to kill us. You know, just good technical topics for CTOs. And at the very end you'll have some delicious stuff about uh, how Lena created some software hand coding some neural networks and eventually got that software OEM in the United States. And lest we forget, she's the CTO and co founder of Good Face Project, a company developing chemistry AI for personal care. Lena, welcome to the CTO podcast.

Speaker B: Thank you, Etienne. Thank you for having me.

Speaker A: Absolutely. What have, what have you been up to today? Ctoing the Good Face Project.

Speaker B: Yeah, mostly typical Tuesday. I don't know for you, my Tuesday are very packed. I don't know, for some reason it's one of the most packed days of the week.

Speaker A: I never know when to do that packed day. I mean I know a lot of days are packed, but I never know should it be the Monday, the Tuesday, Should I. You know, I have a friend who says he doesn't work on Wednesdays. How do you break up your week to stay sane?

Speaker B: I try to pick up as much as possible, uh, in the first couple of days and I think that naturally some things are discussed on Monday and then you act ah, on them on Tuesday and so that's how it goes. Hopefully it's done by Wednesday, but you know, then sometimes it's already Thursday and you are still in the same mood. So you know, depends on the week.

Speaker A: Yes, um, I, for a long time now I have front loaded my Mondays and Tuesdays and then most weeks my Wednesdays and Thursdays and Fridays will be really nice and open. My only problem then is that I feel um, exhausted after the two days. Like I feel like I did a whole week in two days. And then I'm so sort of shell shocked of all the information that was downloaded, all the things we have to sync up, all the things that we have to make Decisions on all the please explain situations of why things aren't the way they need to be. And uh, we're just sort of stuck in that endless loop.

Speaker B: Mhm. And then Monday feels like Thursday, it's not yet Friday, but you all already are exhausted.

Speaker A: Wow.

Speaker B: Yep, we're in the same.

Speaker A: You are the CTO co founder of the Good Face Project. Do you want to share? We had the privilege of having you on our AI panel, uh, on the CTO podcast a few weeks ago. So I get you all to myself now. So do you want to just share for our audience what maybe when the Good Face Project was started and what problems you're solving for whom?

Speaker B: So um, we started working in 2018. That's when I met my co founder Eva. And um, basically from that time, it's already six years since uh, we first discussed the problem of ingredients. And um, the idea came from Eva's research on um, ingredients in personal care. And pretty much at that point nobody knew what is in your personal care, what are you putting on your body? And since your skin is one of the biggest organs, it's actually the biggest organ. Everything is absorbed. What does it mean that you put something on your skin? Right. I never thought about it. I was pretty much never concerned about that. Frankly if it's not a toxic pain then it's fine. But um, turned out that ah, after like very brief research I uh, had to throw away all of my sunscreens that I was using at that point. Like literally some of them were unopened and I just took them and threw away uh, some of them had um, carcinogens that would be activated by sun. So you are putting on sunscreen and it becomes a carcinogen when you are out in the sun.

Speaker A: Really?

Speaker B: Yeah. So it was very bad.

Speaker A: Goodness gracious.

Speaker B: And uh, we started entangling uh, this and um, trying to figure out what is actually used in products, uh, on the market. Basically to figure out what is used in products on the market, what is better for a human being. We ended up with a problem uh, of we need to know everything about all the ingredients and everything about all the products on the market. And you know, it's not a small problem to tackle. So we started moving on and the project evolved from a website app where you could find personalized uh, routines and uh, personal uh, care products that were non toxic for yourself and for your family. And then we slowly but surely moved um, into formulation of products. And at this point we're working with regulatory and formulation um, basically uh, while doing that and that would be like you know, B2B part the engine, the core behind that was evolving. And uh, first it was like a very simple set of data. And then it grew, it grew and evolved into self cleaning, self organizing. Um, and at this point we have a very complex system that ah, um, uh, is basically um, containing knowledge of over 160 millions of ingredients. Uh, and um, we know information about their toxicity, their structure, their functions, how to use those ingredients in uh, products, uh, how reactions are happening. And that all uh, feeds information into the platform that is used by um, chemists, uh, by R D departments, uh to develop new products specifically for now for personal care. So we are serving that vertical and uh, the core right now is uh, pretty much self organizing, cleaning, helping us a lot. So we uh, built a system that can generate information about ingredients that is yet not known. Um, so it is generative AI where you can generate information for um, toxicity, carcinogenicity, mutagenicity, uh, all that bad stuff that is yet unknown, uh, because there was not enough scientific research. And uh, some ingredients will probably never be researched because they're, you know, they're not broadly used, they're just not very popular. But they're still, they can be in your food, they can be in your personal care, in, I don't know, in paints.

Speaker A: What about all the plastic molecules we're supposedly ingesting?

Speaker B: That too, right? So pretty much anything that has a specific molecule structure can be analyzed and uh, the system can be projecting information about their toxicity and qualities.

Speaker A: So uh, 2018, obviously the uh, GPT hype was not, did not exist. Did you just go with traditional machine learning? Data sciencey? Back then it was called Data Science, Big data.

Speaker B: And you can change the names and I think that we will change names many many more times until like we arrive into something that is universal, unique. Um, but um, I think it's a moving target. AI is a moving target. Humanity. Whenever we're getting to the next point, right? Like Deep Blue, not an AI anymore. Like, oh, we passed the Turing Test. No, that's not enough. So you know, it's like never ever we're going to reach the point even I think if we have a robot that's talking to you and fully functional and can do like plumbing, people will be say like, well it's just a robot, right?

Speaker A: Maybe AI is the moniker that or the banner we put over the mysterious, you know, intelligence. We don't really know what that is. And we're going to put artificial so that it's man made and then general, because we don't really know. So let's just say AI. I, I, I love how, I mean this is such a cliche by now, but I love how everything is the AI now. Oh, we, you, you put someone's name in a form and then the AI will generate a PDF with your name on it. It's like, no, I don't think that's AI. I think that's um, uh, a model and a view and a controller.

Speaker B: That just, that's the hype, right? That's the hype that we discussed last year. I think it continued and the result was the crazy amount of companies that are, I don't know, doing PDF reading and some other stuff and they're like saying that they're AI in this wave of companies that will be rising from this and then they will probably most likely disappear very quickly, replaced by new companies or, you know, bigger companies actually doing the same thing. We'll see this a lot.

Speaker A: So are you uh, you're a venture backed company?

Speaker B: Yeah, we do have vc.

Speaker A: And so have uh, you found that raising money has become exponentially easier?

Speaker B: Well, we did not raise since the hype started actually.

Speaker A: Hey Lena, you should, you should raise.

Speaker B: I mean, yeah, we're kind of, you know, old fashioned not running to raise all the time. Um, but um, I think that VCs were like the one, the wise ones were very cautious, the ones that actually know what is going on. They very diligently uh, looked out and checked and basically tried to figure out is there actually something behind that. Uh, and you know, many companies arised and they were just coming with an idea. Right? We have an idea. We don't have a product, we just have a deck. And right away they wanted like, you know, very, very high valuation. So I think it was like they were riding the peak of the wave of the hype. Some of them definitely succeeded, but I think that majority did not.

Speaker A: Yeah, ah, I, I hear you. Um, as you were talking, I was you know, just thinking if, if I were a VC and you were explaining this technology to me, you're obviously talking to ctos. But help me just so I don't spin my brain on this for the rest of our conversation. Um, you clearly use a generative model to calculate, is it like the multiple permutations of molecular bindings and given, ah, or are you reading a list of ingredients and then you do stuff like can you just.

Speaker B: And that and many, many, many more things. So I guess that um, the thing about the real AI is And I'm talking to, like, CTO to cto.

Speaker A: Yes, yes, yes.

Speaker B: So the real thing about this is that it's just not one module, just multiple ones. Right? Um, we can see, we can hear, we can speak, we can think, we can balance, right? And we have different parts of our brain. We don't expect our brain to be put mixed in a blender and still be capable of doing all that stuff. And what we're doing right now, we are creating a blended brain. And, you know, that's the, that's the outcome, right? So, like, the results can be very weird, but that's because this, this parts are all mixed together. And in LLMs, like, it's like, it's not really what we need. At the end of the day, LLMs will be just a single part of a bigger chain. Multiple, multiple modules. And I'm pretty sure you heard this many times from CTOs and you had a question about how would I describe it to vcs?

Speaker A: No, I think, uh, as you were talking about, as we were talking about raising, I was just thinking if I'm sitting in the presentation, I would like to just know the shortest path in, like using the Good Face Project, Do I. Am I scanning a list of ingredients and then you tell me if there's toxic elements in there? Um, every time, every time I give you an example, you're like, yes, that's what we do. Yes, yes, yes, yes. You know, uh, the other day I was. I went to a baseball game with my son, and we like to buy Cracker Jacks and, you know, the little caramelized popcorn for all my international friends, Cracker Jacks. Um, and I had a moment, and I had this moment actually quite a lot, but it was pretty vivid on, On Sunday where I, uh, you know, I opened the packet and I looked inside and I saw these perfectly formed popcorn thingies. And you know, of course in my brain I'm like, how did they get it so perfect? You know? And then my brain goes to modified foods and processes, and then it's like, is this thing that I'm having actually a popped kernel or is this just look, you know, like McDonald's fries just look like French fries, but they're not. Um, this episode is sponsored by McDonald's now.

Speaker B: Um. Ah.

Speaker A: And I was just thinking to myself, the manufacturing process, the packaging process, the chemical process, everything that went into that little thing that I was about to trust to put in my mouth and, and potentially suffer the chemist chemical consequences beyond what I could know in the short term and Definitely don't know about in the long term, you know. So I know a lot of you are listening to this thinking, okay, well, clearly, you know, that's the supply. That's the food chain. Like that's the supply chain for food. It's not a big deal. But I was just marveling at my trust and my son's early life trust that there's a packet and the whole development process to get to that point so that I could stuff my face with what this looks like popcorn. Um, kind of when you talked about the toxicity levels and scanning, what's really going on in there just, um, gave me pause.

Speaker B: Yeah, it's pretty scary if you start thinking more about this because, you know, all the supplements, all the like, personal care chemicals in your house, everything can contain a lot of not disclosed ingredients actually. Uh, and, um, definitely FDA is not doing best job on regulating them and keeping people safe. So, um, I must admit that, um, I really like, uh, products from Europe because I know they are way more strict and so whenever I can, I, I really prefer buying stuff from.

Speaker A: I. I know it's pretty incredible. Um, when I see the CE mark on anything, I just, I don't know what the one is for food, but I generally just feel, wow, okay. You know, these guys, I mean, of course they feel like they're over the top, but why wouldn't you be over the top when it comes to what I do to my body?

Speaker B: Yeah. And especially when it comes to kids,

Speaker A: you know? Yeah. This is, uh, you're activating my parents side now. But America's got to be the only country on the planet that is actively killing its people with its food.

Speaker B: Food products, and you are not even considering yet pesticides that go into the ground like all the time.

Speaker A: So, um, why would we, why are we okay with this? I don't understand. I do not understand. Cool. This was a great, uh, CTO podcast. So, ah, uh, speaking of, um, uh, parenting, uh, I see one of the most important books you've read is on parenting.

Speaker B: Yeah.

Speaker A: You want to share a bit about that?

Speaker B: Um, well, I have two kids and I, uh, try to be a good parent at the same time as being entrepreneur and cto and you know, sometimes it's tricky. My kids are very different. I have a girl and a boy and they're like in and Yang. So I get to experience the whole spectrum and sometimes I question my methods. So you know how you sometimes.

Speaker A: You question your methods. So many great titles for this podcast. Brain in a Blender, Questioning my Parental Methods. Wow.

Speaker B: Cut it out.

Speaker A: The government's trying to kill you. We're acing this. This is great.

Speaker B: Okay, so like, remember the regrets?

Speaker A: Yes. No, I don't have any regrets. You might have regrets, but I don't have any.

Speaker B: Yeah, it's like, you know, you have work skills, you have parenting skills and you know, you can always think if you're doing things good on um, either side. For example, it was very important for me to know if um, the like, if I'm over reacting with screen time. Right. Maybe it's fine. While like I prefer to limit screen time. While you know, some people banned completely screen, others are like not banning either. Like what do I do? And like where is like so you know, I just like to from time to time check in if I'm doing the right thing.

Speaker A: Yeah, it's a, it's a big one.

Speaker B: Um, how about you? Like, are you revising your parenting skills?

Speaker A: I, I, I, I empathize with you greatly about um, my parenting skills. We have a boy and a girl and a girl and um, two things when it comes to screen time. This worked really well. When they were in elementary school we taught them how to regulate. So we, the regulate is a big word in our house. So I tried not to ban screen time, I tried not to have timers and I tried not to um, you know, have completely unsupervised screen time. But the thing that I was trying to teach my kids was just how to regulate. Learn what's happening in your body. Be observant. You know, daddy, um, would be on the phone all day as well if I could. I can scroll TikTok for five hours if I have to or if I don't have to. What I actually try and teach my kids is that the addiction that they feel for screen time is exactly the same addiction that I feel as a 52 year old M does. Not gonna, it's not gonna change. What is gonna change though is if you have the ability to regulate. And so this worked really well when they were in elementary school and it also worked really well when they were uh, it works differently for each one of them. So I have three and so you know, each child like you said yin and yang and yin or something. But, but um, I did notice that the, the kids would respond to it differently. Uh, so yes, all of a sudden I would see the kids playing outside. This was miraculous. Like we would not tell them to stop playing with their iPads. We bought them all iPads. They, I am, I am technology positive with my kids. Although I do realize that the dopamine levels and the brain can't handle what's being thrown at them. So I am aware of that. But when they were in elementary school, they were able to, like, I would just see them go outside to the trampoline, to the this. Or the kids would talk to each other and say, okay, I will tell you when we're done. That was pretty amazing. That was the regulate phase. Um, they're in middle school and high school now, and it's, that's not working. Okay. So they, the value of, and I mean, you have covert kids, I have covert kids. But the value to them, the, the interpersonal relationships are now via phones and games and the online world. Um, they have constructs in their brains of their friendships that is in no way validated by real world experience. Although now my son does go to school with some of them and you know, they, they mess around and all that. And I can tell that little humans still need relationships, physical relationships. So I, I, I love that. And, um, but the journey is very difficult. Um, but I think telling my children that I have the same problem they have has been a wonderful normalizer so that we can all know. Hey, so one rule we do have, which is a pretty hard rule, which is very, very difficult, is on Sundays. It's no tech Sundays. And I am religious about that. Like, I, even if they weep or if they give me the best reason under the sun, I say I'm really sorry, but I know you have to look that thing up. I know you have to want to listen to that song. I know you're just trying to look something up quickly, but I just like go totally Luddite with them on Sundays, which is very interesting behavior that comes out on Sundays.

Speaker B: I really want to read your next book on parenting.

Speaker A: Wow, that's a good idea. Yes. The other thing I do, Lena, um, which has been amazing, talking about the collapse of parenting, which is the book you refer to, and, um, constantly wanting to learn from others on what to do better is I do regular. And I don't want to sound like I do this on a certain day at a certain time. I'm not a process person at all, so don't listen to me around process. But I will regularly check in with my children and do a retrospective with them. Do you do that?

Speaker B: Yeah, it's like check in, like stand up.

Speaker A: I do, uh, what did daddy do great. Where did daddy mess up? And what can daddy do better? I tell you, Lena, that was a game changer for us. Because, you know, what happened was my daughter, my middle daughter, who's a little more anxious and on the more sort of emotions on her sleeve kind of person started, started telling me, because look, she is. They are still young enough to remember their early memories. Like you and I can't remember when we were six. Right. But a ten year old can remember when they were six. So. So my daughter would start divulging to me, um, very scary thoughts that she had when I was angry or when I raised my voice or when I was displeased. And I tell you what, it has been nerve wracking to

Speaker B: get this feedback,

Speaker A: get that feedback, because I'm of course very sad because I see my own inefficiencies and inadequacies as a young parent not knowing what to do. And you like, yell and you raise your voice. I said to my kids every day that every retrospective I say to them, if I could do parenting over again, I would never, never ever raise my voice, ever. There's not a single benefit I think that kids get from that. So that's just a little something. But, um, yeah, so I would get some pretty deep retrospectives, Lena. Deep.

Speaker B: Yeah. I'm at the same page. I'm also getting feedback. And, uh, you know, I try to become better and evolve. That's just a path. I try to forgive myself, then move on, just be better.

Speaker A: Yeah. My wife said to me sometimes when I think I have irreparably damaged my children, she says, as long as they keep coming back, you're good.

Speaker B: Still counts.

Speaker A: As long, as long as they feel like restoration is accessible to them, then you haven't damaged them beyond repair. I know, um, as parents, we often think, oh, my goodness, that I just damage my kid. And, you know, so, So I do rely on them, like you said, forgive ourselves. I do rely on their forgiveness as well. And I remember my dad, my dad didn't have the tools back then, and parents didn't do that. But my dad never really apologized to me as a, as a young kid.

Speaker B: It was not a thing, I guess before you just, you know, you survived and that counts as success.

Speaker A: You're alive. Uh, and I will, uh, add a little asterisk. I don't remember the apologies my dad may have, but I don't remember them. But I do, I do access that a lot, is how can I ask forgiveness? And it's kind of incredible. It's an incredible moment when you can just say, I was impatient, I am sorry, I, uh, I was completely unreasonable or I shouldn't have said that. Or, you know, when my, my natural instinct is to push through with perfect parenting. Well, I said that because this is what I meant. Like. No, dude, you said that. You shouldn't have said that. Just move on.

Speaker B: Yeah. And I guess it's just this, um, phase of parenting. Right. Like if you, we look at our parents and their parents, the way kids were raised, the approach was very different. Right. So people were not very, you know, cognizant of emotions. You don't have emotions.

Speaker A: Yeah. You didn't have a relationship. What's this thing called a relationship between a father and a son? That, that's. I, I see my dad and his dad, his dad, uh, my granddad, who passed away about 15 years ago. 10 years ago.

Speaker B: Ah.

Speaker A: He was a train driver. And there was no access to emotions there. There was just a smack through the face if something wasn't right.

Speaker B: But that was very fast. Right.

Speaker A: So speaking of, uh, your parents. So you were raised in Ukraine?

Speaker B: Yes. Mhm.

Speaker A: Mm.

Speaker B: But, uh, you know, in general, I think that parenting, uh, did change across the whole globe. So, um. And I'm just guessing that you are from Generation X. Yeah. So like, this generation is specifically, uh, interesting because I like, I learned a lot about, you know, my husband is also from the same, the same generation. Like the, the universal experience is actually the same across the whole globe. Pretty much the same approach. Same like, you know, for example, when My husband was 2 years old, he would be let out in a city, like it's small, it was a small town. Okay. But they would let their kids out, uh, and just go like two year old kid. Just, you know.

Speaker A: Two year old.

Speaker B: Yeah, two. Like, I, I don't know.

Speaker A: I can't let my 15 year old walk to the bus. Bus stop, which is just down the road from here. What happened? Are you ready to elevate your team's performance and your own career as CTO? Introducing the 7ctos Growth Program. With 7ctos, you'll join a supportive community of tech leaders dedicated to helping you reach new heights. Our group coaching program, led by experienced CTOs, focuses on the proven CTO levels framework. Assess your team processes and technology challenges with our, uh, easy to complete assessments. And then dive deep into monthly coaching calls where you'll get unstuck and gain insights on specific topics. Craft actionable plans tailored to your needs, guided by expert advice and playbooks designed to drive results. Plus gain access to our exclusive CTO community for ongoing support and collaboration. Sign up now and start your group coaching Adventure for only $300 per month or $3,000 per year. Your membership includes priority access to all seven CTOs events and activities. Don't miss out on the opportunity to join a global community of experienced CTOs, attend virtual events and participate in confidential group coaching sessions. Visit sevenctos.com growth and level up your leadership skills today. So did you. Were you raised in communist Soviet Union?

Speaker B: Yep. I was born during that era, so I, I saw different, uh, phases of, uh, the country. I was very little when, you know, when was still existing. Then I saw the collapse, what happened after that. So I guess, um, that background, uh, shaped a lot of, um, you know, how I now go through things, how I work. Right. So. And my, uh, co, uh, founder, she's from Bulgaria, and although, uh, her country did not collapse, it also went through crazy, uh, era of hyperinflations and everything. So she saw some tough stuff. So when we, uh, you know, found each other here in San Diego, I guess it was match made in heaven. It was like two people just right away, uh, able to work, able to, you know, find language. It was some magic. And I guess it was brought up with our background.

Speaker A: Yeah. I, um. Didn't you do like a founder match thing?

Speaker B: Founder, Founders Institute. So I did it in, uh, 2000, um, 16, I guess. 15. 16. And she did it in 2018. So that's how we met at the end of the day.

Speaker A: Amazing. And you know, uh, your. I think when we spoke last, you mentioned your dev team is in Ukraine.

Speaker B: Yes, most of them are in Ukraine, some of them are in Europe. But so.

Speaker A: Yeah. Ah. Are you. Are you comfortable with kind of sharing how you navigated the last. How many years has it been now? I mean, it's,

Speaker B: I think all the time. You mean since the war began?

Speaker A: Yeah.

Speaker B: Um, so, you know, it's like moving up to three years, like it's to end. I must admit that we went through several phases. I guess the, uh, you know, at the beginning it was, um, super, super stressful. Super. Well, it's always super stressful, but, um, you know, the shock and um, pretty much inability to predict what's coming next, like it was insane. And uh, nobody knew what's going on in the first couple of weeks. So now we are kind of in this. It's still war. It's still super bad. You know, my mom is in Kiev and I, like, on daily basis I see how that new bomb was dropped there. Uh, but, um.

Speaker A: So are you saying that around where your mother is?

Speaker B: Yeah, well, you know, all the country has been, um, yeah, the whole country. So it's every day, like literally every night they send out a lot of drones, sometimes rockets, and it's just happening all the time. I know it's not on use, but it's like literally every day. Um.

Speaker A: Wow, man.

Speaker B: So, um, at the beginning, you know, at the stage when nobody, uh, knew what's going to happen tomorrow, um, frankly we expected that our team would be stagnated because that's what you expect from people in this situation. Right. You know, they would be running, saving themselves in shock. And um, you know, I had some very, um, memorable calls with them when I would talk to them and then would say, I need to go to a shelter because like bombing is happening. But you know, the most shocking thing from that situation was that they actually pushed to push new updates. So, you know, Russia is invading like every, Everything is on fire, everything's in chaos and these people are pushing to prod. And that's when I was like, okay, if they're doing that, wow, um, um, we can go for anything. Um, it was insane. But uh, you know, somehow even to this stage of insanity, people get used to, uh. It was tough. Also another phase was very tough when, um, Russia attacked a lot of, um, power plants. And uh.

Speaker A: Yeah, so that, that's targeting the infrastructure.

Speaker B: Yeah. And that resulted in no electricity. And how do you, you know, how do you work coding when you don't have electricity? So these people got themselves, um, where you put some gas into it and it produces electricity and they got themselves starlinks and other ways of connecting and they kept working. So, you know, that's just incredible amount of resilience in these people, uh, where you don't even expect people to be able to do that. So. Yeah. And uh, it just continues.

Speaker A: Yeah. I mean, Lena, that's um, that's, that's crazy. And I appreciate you sharing, you know, because obviously when a war carries on for so long, uh, people can sort of get used to it. The, the, the west or the, you know, people who are not directly affected by it. And also clearly the media, you know, it doesn't fit the news cycle anymore, so we don't get the, the real news. I, I remember there was a 7ctos member who, right when the war broke out, um, he was a cto, he was based in Ukraine and he was, I think, in one of the southern towns closer to Crimea.

Speaker B: Mhm.

Speaker A: And I would be on slack with him and then he would say, there are two helicopters hovering above my house right now. I have no idea who they are, I have no idea what's about to happen. And like running around and, and you know, and then, you know, it could have been, it could have been Ukraine helicopters, it could have been Russian whatever, but, but um, to live under that kind of terror. Normalized or not normalized, but the IO happening through Slack, like I'm having a chat interface, I'm sitting in my lounge, comfortable, sipping something and then this guy is running away from hovering helicopters.

Speaker B: It's just insane.

Speaker A: It's really the humans, the human brain finds it hard to sort of wrap one's head around it. And so when I hear you speak and you relay these actual real world stories, I think for me and for everybody listening to this, it's important, very important for us to hear that.

Speaker B: I think you're completely right. The news is just not about this. And uh, it kind of almost expected it, right. Everybody at the beginning was shocked, but it was obvious that you know, everybody just cannot be, be getting these news all the time. So you know, the politics, people are tired of the one and the same story about one in the same country. So you know, it stops being used. And um, it was predictable that at some point Western, uh, society will be just, you know, this less interested. Um, that's a, uh, very sad. On the other hand, I must admit that Ukraine itself is just standing firm, like as firm as possible in this situation. You can always wish it was better, right? Some better decisions were made, some better preparations, um, better governance. But well, we have what we have. So we just try to make the best out of that. For me, I try to like part of um, what we're doing is actually, you know, keeping, keeping working with these people. It's like very important for me, like some sort of support. I don't know if that makes sense but.

Speaker A: Well, and I imagine I, I can only imagine that in a surreal war zone environment for your team, they need, they need that connection to normalcy, to tech startup. You know, like there's, I can imagine that that's something that is maybe gives them hope and maybe as a co founder and as, as the cto, you have a very important role to play in giving them that normalcy and that hope.

Speaker B: I think uh, it makes sense like uh, just this normal part of normal life where you have job and you work at something and you know that that is already something that can carry you through. Um, yeah. So um, I'm sure that you had many stories from different CTOs from different parts of the world.

Speaker A: Does it affect your ability to be um, firm with them or upset or real or. You said you were going to ship it last week and you didn't do it. Or like

Speaker B: when we need to before and we're from rights. I must admit that our team is um, very strong and um, they like what they are doing which is already very cool. And they just, they just really do their best. I know and I watch, we do like watch the time we do have, you know, stand ups and like all the reporting. But also I can see the attitude and uh, that also counts.

Speaker A: Yeah, yeah.

Speaker B: Where you, you know, sometimes you would have a uh, developer to whom you are giving tasks and they're just completing the task.

Speaker A: Yes.

Speaker B: Versus uh, most of our team. If you give them the task they would just, you know, explode with additional ideas and like. No, like do it that way. I told you. Okay. Okay.

Speaker A: How, how large is your team?

Speaker B: Uh, 27 people now.

Speaker A: So specifically in engineering.

Speaker B: Yeah, like in this department.

Speaker A: Got it, got it. Okay. The way that you view AI, I know, um, I know you're the CTO of a company that is pushing AI into production. AI into production. But can you share with the, with, with us a little bit what literally what your tech stack is and kind of what you're, what are you coding and what are you pushing and what are you, where are you pulling, pulling your models from or deploying? Just a little bit of a geek speak on Goodface's um, Tech Stack.

Speaker B: Um, so I guess I uh, was discussing it a little bit last time we had a podcast. So uh, um, I would split up uh, our tech stack into you know, what's publicly available and what's you know, was built specifically for us, uh, by us. Um, and I'm pretty sure that the most interesting part right now it's not like oh, we're using Kubernetes or like I don't know, Aurora, but people would be interested what models are we using? Right. So um, and um, the answer is we um, actually are using, trained to use the latest ones. So. As simple as that. Um, because uh, you can go and find so many models for LLMs. Um, uh, for example, at some point, uh, Mistral, Mixtral were amazing, right. You could compare them to chatgpt and uh, Llama and others. So they were probably at the top. Now uh, if you prefer to build everything locally and not use ChatGPT APIs, although they promise that they are not using your data in any way or storing, I'm pretty sure uh, there should be some imprint.

Speaker A: It's just, you know how can they not be?

Speaker B: Yeah, I don't know how they can not be. So it's just in the core of the technology. So we prefer to do uh, most of our um, calculations, um, locally. So we have um, Lamont, the latest Llama is very, very good. The smaller one, I don't remember. The size, like 8 billion. I guess. Uh, that is pretty much enough for us to work with the texts that we are right now producing for our system. But uh, frankly our system is m. Mostly not about text. It's mostly about precise data. And precise data is um, uh, pretty much either deterministic or probabilistic, where you know, the like exact probabilities. Right. So you know, you can have a cutoff. And uh, that's why uh, uh, there is a lot of um, proprietary stuff that we built uh, for so many things. For understanding um, the molecule structures, for predicting different, different types of toxicities, for predicting reactions, for formulating a product, for saying um, what would be the concentration of an ingredient in a product, for predicting what a human being would be saying about the products. So there are so many things that we're between. And yes, that part is using LLMs,

Speaker A: the stuff that the humans need to read. Yes.

Speaker B: Right. So whatever is not precise to the, you know, to the numbers, it can be done by LLMs. But still you need to double check it, like to cover it with as many agents as possible and you know.

Speaker A: What did you say? Cover it with as many agents? Yeah, testing agents.

Speaker B: That too, yeah. So you, you have results. You need to go back and check the results. So you need.

Speaker A: Absolutely.

Speaker B: Source.

Speaker A: I mean, how do you, how do you even check whether the monstrosity that you combined in molecularly is even a thing?

Speaker B: That's very tough. And that comes from knowing, um, many, many, many structures of molecules and many, many facts about them. And those facts need to be precise, they need to come with a source. So you know, whatever is a fact in our system needs to be supported by multiple sources.

Speaker A: Got it.

Speaker B: And uh, this means that uh, for example, if we have a fact, but that fact has just one source, the probability that that is a fact is very low. And then, um, how far can we take it in using for further predicting something?

Speaker A: You know, so, so, so a fact could typically be. Here's a written chemical composition. Here is its toxicity levels. Here is. It's this and that and the weight and the this and the. Is it something like that?

Speaker B: Yes, that. Um, then, um, all the papers that can be attached rates interesting. Some scientific papers can be stating that it's a carcinogen or you know, it might not be, uh, supported by a paper, but um, some company reported it. So summary of that basically uh, is capable of moving something from just a possibility to a prediction with some threshold. And then it's on you whether you know, the threshold is 80%, 90%, 95%. Where is the truth? And very often in our system, uh, we try, you know, to basically predict with the highest probability, but at the end of the day taking information into account and acting upon it is uh, uh, you know, it depends on our customers. Right. So we can produce information, uh, we can show them where it comes from. But then whether it goes into some action that already depends on the human.

Speaker A: Interesting. So that, so you're. That's very interesting. So I think in back of my mind I was thinking you may have all the rules of chemistry composition and then you, from first principles you. And I'm sure you're going to say yeah, we do that as well. But um, but it's interesting to come the other direction where you now go into the wild and scientific papers and science and you go look at everything out there and now you build your models, you know, that contains all this um, additional information about the compositions.

Speaker B: And you are correct. Um, we do have like, you know, basically building blocks of molecules and how they combine into each other. So that also is a thing that is used in some of the, you know, pieces. Yeah, um, it's like sometimes it's very hard to determine and you know, some, some data for example, um, uh, some isomers, uh, which is a molecule. Um, it's like mirror isomers. So basically there's just one tiny element can be rotating in another direction in the molecule and the whole molecule can have other functions just because of that. This is where data stops and that can only be pretty um, much detected by some scientific devices. Right. So there is a threshold where the data stops and I don't know, almost quantum physics begin.

Speaker A: Wow.

Speaker B: But uh, until then we try to figure out facts.

Speaker A: So your typical customers are manufacturers, um,

Speaker B: regulators, um, pretty much anyone who is working in the personal care industry. And uh, it begins with brands. Yes, it's manufacturers, it's retailer. Uh, so basically uh, different kinds of uh, customers can find different value in the information in managing their portfolio and see, seeing what is, what is actually going on. And uh, one of the big things that we're providing is actually regulatory assessment of products. And uh, that is a uh, very hard task um, because you know regulations keep changing and there are hundreds of them. It's moving targets. It has to be very precise or like it just doesn't matter. So that uh, is also important. And uh, like you said, regulatory bodies are also using us. I wish we could work with FDA and you know, help them organizing their data. I just wish. Unfortunately, uh, you know, that's probably not the biggest focus of FDA currently.

Speaker A: The other day I was, uh, my, my wife had some uh, um, acne medication and creams. Uh, and in my mind I was looking at this and I was thinking if I were uh, I wonder what services exist to help this company, startup go from idea to product in people's hands. So if today I decide I want to create a capsule that will hydrate me to 100% and all I got to do is take one of these once a day and I am now completely hydrated.

Speaker B: Okay. Okay, sounds good.

Speaker A: And I run my TikTok branding and I ace it. And people are like, dude, this is incredible. Now I'm sure, and I don't know this for a fact, but I'm sure that through Chinese manufacturers and services out there, I can create a pull capsule. Oh, do you want it to be 5 millimeters? 8? Yeah, I want it to be blue on the ones you. Does you basically design the thing and. Oh, do you want, what do you want in it? Well, do you have a drop down list of things I can put in it? And, and then, oh, the packaging and then the distribution. Uh, all that stuff seems to be done for you in many, many, many aspects. How do I know that? Whatever my kids are putting on their faces. I mean, I guess you got brand recognition again. We're back to crackerjacks. But uh, a. Is it a regulated environment? So can I just slap on. Can I just put pills in a bottle and start shipping it worldwide? Or is it, is it. Does the FDA approve that stuff?

Speaker B: So, uh, basically just this summer, FDA started requiring companies that are making I guess more than 5 million to post information about their products into FDA storage. And that's uh, it.

Speaker A: Wow. Wow. So if you're, if you're some random janky place that makes a hundred thousand a year because you put stuff in a bottle and you have people drink it and you run a TikTok campaign every. Go for it.

Speaker B: It's a bit different with food and personal care. So if we're discussing.

Speaker A: Let's talk about personal care. Actually, uh, I don't want to. I realize food is different, but, but personal care. Can this. Does uh, swallowing a capsule include.

Speaker B: Is that considered food yeah, that would be supplements.

Speaker A: So uh, and that would be considered, that would be highly regulated.

Speaker B: Well, it's not very highly regulated, but it will be closer to food and medications. It depends on what goes into your capsule. Right. Uh, in personal care the regulation was pretty much non existent until like I said the summer. Now, um, there are uh, there were like 14, I guess it was 14. 14 ingredients that were regular restricted by FDA 80 years ago, like several years ago. That was the story basically 14 ingredients 80 years ago. So they started doing something in this direction, starting. But you can see there is California prohibited list. There is some another state prohibited list. So since FDA is not doing excessively good job on regulating ingredients, some states and California of course typically goes first start regulating some ingredients. And you know, uh, California Prop 65, you must have seen it, right? Heard about it?

Speaker A: No, I don't really watch the news.

Speaker B: You can see uh, sometimes on some objects that it contains carcinogens.

Speaker A: Oh yes, at the gas station. And is that Prop 65? Okay.

Speaker B: Uh, so this summer uh, they started requiring posting this list of ingredients. Actually if you um, misrepresent contents of your products and FDA finds out, you are in trouble. So uh, to do that they would have to uh, you know, decompose your uh, products, analyze it for a list of ingredients, detect whether there are some impurities in that product or you know, some ingredients that should not be there, then they will come after you. But it, if they don't do that research, that's it. And you know how many products were uh, analyzed?

Speaker A: Um, and they probably. It's like IRS audits. Like you, they just can't get to everybody.

Speaker B: Yeah, absolutely. And uh, such uh, analysis it requires uh, an expensive equipment. It would be like, I don't know, it might estimate up to $30,000 per product. So I imagine FDA will not do that.

Speaker A: A perfect opportunity to sort of deconstruct, I suppose. I don't know. I don't know if you can.

Speaker B: Well you can use to some extent spectrography, but again to some extent. Right. So you will detect some ingredients using that, but you will never detect whether you're using orange oil or lemon oil. Right, so. Or was it, I don't know, some cabbage oil? They will not ever find out what is in your product from the plant perspective.

Speaker A: Yeah, I suppose it's just up to the consumer then to just be careful.

Speaker B: It's very encouraging, isn't it?

Speaker A: I just want someone to care about me and protect me.

Speaker B: Okay, uh, and here we end the conversation.

Speaker A: If only there was a parent who was taking care of me. I think this is true in America especially is, and both of us are immigrants. But um, it's really shocking how in America you really, it is every person for themselves, even the healthcare system, everything. It's just, you gotta just figure it out man. Google it.

Speaker B: And then you go to the doctor and you would say I googled it. And then they.

Speaker A: Yes. And I hate that. I hate that. Well, uh, maybe, maybe tell, tell me a little as we end. Um, you know, I know you're passionate about AI. I saw also that you uh, started your work at the Ukrainian Space Research Institute. Can you tell me about that just for a sec?

Speaker B: Mhm. Uh, well I guess it was always a passion of mine, space and everything related to it. So I guess if I was born in United States I would be right now somewhere in NASA or SpaceX and would be pretty happy with that. Um, in Ukraine. Um, Ukraine has a very good history of uh, space um, industry because there were some rockets that were developed in Ukrainian uh, research institutes. And um, you know it was always a dream of mine to uh, work with space. So um, for a while I was working there in basically um, at that point we were doing the image recognition uh, and reading images from pictures that were made from space. Now it's a big thing and you know like all the companies that are doing that are successful they're like making um, recognition of images and pictures made from space. Uh, we were doing uh, also spectral analysis of areas and uh, in that uh, aspect we were detecting areas that were dried out, very prone to fires or um, there was a contamination of any kind. So basically there was a ah, lot of research done at that era. I think that as usual our scientific researchers are very much ahead of the curve of people actually being ready to something. Uh so for example I know that um, in the building next to ours there was a company that would um. Well it was not a company. It was like actually it was young people doing that research. They would uh, create a 3D, uh image from three pictures. I mean it was 2003.

Speaker A: Amazing.

Speaker B: So uh, it became a thing somewhere in 2010, 12 I guess. Right. So uh, there was very interesting areas of research and um, I really, really enjoyed working in that space. But um, I must admit that at the same time I started working on my own software. So um, I liked uh, data mining and that's when I uh, uh, coded my first uh software for data mining. So I, at that point I invented several algorithms for data mining that would find hidden Detections in, uh, data. Uh, so you could use, um, well, data in a format table, Excel, spreadsheet, you know, database. And you would, um, basically point the system to the, uh, fields that you are interested in. And, uh, you would want to predict sound outcome, some other column, some other field. And the system would create a set of rules based on data, combine it, give it statistics, and then, um, would train, um, neural network. And you had to make neural network. At that point. You would have to code it. So it would train that network. And then you could use this data to predict even further. Um, so that software was, uh, sold in the United States. It was OEM by a company here in the United States.

Speaker A: Wow.

Speaker B: Um, I remember my first customer was from American University. You know, like, it was 2000, probably four. And then somebody from a huge university in the United States is buying your software. I was like, oh, my God, it was amazing.

Speaker A: That's incredible.

Speaker B: Yeah. So I was working in this area for a long time, and, um, I think this precedes current AI. Ah. Where you need data to have an outcome with some probability. Right. So it was another flavor of what we are having now. Um, so I guess it was always there. It's just, you know, evolving. And we're always unhappy.

Speaker A: Yeah. Did you say always unhappy?

Speaker B: Yep. With the results, right? Like.

Speaker A: Oh, yes. Yes. Perfection. Wow. That's amazing. Lena. Uh, that is quite fascinating to detect hidden dependencies between data. Thank you for hanging.

Speaker B: Thank you, Etienne. It was awesome.

Speaker A: That's the show. Check out ctopod.com. stay connected. We love hearing from CTOS. We love hearing from CEOs. Anybody who needs to get their CTO plugged in, check out 7ctos.com. There are membership levels for everybody. So it's never too late to expand your network, nurture your relationships, and please, let's see each other soon, like next week. Cheers.