 Genevieve AI is one of the hottest hobby right now. But what does it look like to be able to become the head of AI for a global company and working on Genevieve AI and also lead an amazing team to achieve greater purposes? And what does it look like to be able to move from Brazil to UK with visa sponsorship and really create an international career? There's a lot of new opportunities that Nicholas created for himself, for his family. And in this video, our guest speaker, the head of AI, Nicholas is going to share with us how he created international career as the head of AI and it saves the children international. Hey, guys, this is Dr. Nancy Lee, a director of product and feature in Forbes that helped 100 people learn the dream PM job offer in fan companies, a unicorn startup and continue to get promoted as a product leader. In this channel, we cover tech trends and free product management training. Like and subscribe and check out our new video every Tuesday. And today we have the head of AI, Nicholas, saves the children international in UK to share with us how he created international career in this very hot tech topic right now. So now let's welcome Nicholas. How are you, Nicholas? Thank you for joining the show with us. Oh, so thank you for inviting me, Nancy. I'm very happy to be here. Awesome, it's such an honor to have you on our show. So Nicholas, can you quickly introduce yourself to the audience? Oh, hi, I'm Nicholas and I'm a 10 years of experience product manager in the AI field and digital health. Right now I'm working at the Save the Children International as head of gender TVI, an NGO that impacts more than 50 million children around the world in more than 100 countries. So this is my dream job right now. So I'm very happy to be able to deliver the state of the art technology for good without having to spend so much brainpower on how are we going to make money out of this? Awesome. So Nicholas, and actually I was very inspired by many even things in your career. First of all, it's your immigration journey and second is to actually become the head of AI. So let's talk about your entire journey based on my understanding of your background. Actually you only were educated in Brazil, you're Brazilian, you never worked outside of Brazil. And right now your company sponsors you to move to UK. Can you tell us how did you actually become the head of AI in a UK based company as a Brazilian immigrant? So I was struggling a lot with how to organize, LinkedIn, resume, everything like that. So I was two months looking for opportunities and struggling a lot to get interviews. Then I found out about the PMA and then I was able to organize everything, understood what I was doing wrong. And the key thing was to understand that being a great professional was completely different from being a great candidate. So I needed to understand how to play this game of a candidate, how to bring myself better, how to communicate better. So all of this I learned. And after getting set the core things, I was able to get a lot of invites for interviews, dozens a month. And I was able to choose very carefully where I would want to be. So that's when I recruited from the State of the Children International approach me. And it was very interesting because I was in a lot of interviews before for U.S. And one of the problems I was having is that the recruiter asked it if I was in U.S. already, but I was in Brazil working for a company based in U.S. And another interesting information is that a lot of whole girls had the information that we need someone who sees diversity as something has a power, a strength, as something important, okay. But I found out that most of them don't just write this on paper. So Save the Children International, I'm working one month already and I see the client is in practice. So I'm very happy to have found a company that truly sees diversity as a power. So talking with the recruiter, it was very interesting because I decided to be myself and be more humane with her and show my passion for doing technology for good. And then she understood that I was, they were not looking for a candidate from somewhere outside U.S. So they were looking for the best AIPM candidate from U.S. And they were able to find me. And then in the interviews, the recruiter understood that if I were the best candidate for the role, it doesn't matter where I work. So everything worked out and the institution struggled a lot with bureaucracies to get me to UK so that they can hire me. And I really, really truly appreciate that. So it was not an easy journey, bureaucratically wise, but we were together on that. So I felt really embraced it. Awesome. So right now you're already starting your job and a sponsor visa. I also heard that they're gonna sponsor a whole family to move there as well, right? To UK. Yes, yes. Not the whole family because the cats are the additional things, but the visa process, the travel to get the biometrics, everything. So I'm really glad. This is awesome. Such inspiration. Also the first month of expenses. So I truly appreciate that and that's it. Poor cat. She needs to learn British accent before she can move over. But I'm so glad the company actually discovered that you are the best candidate, regardless where you're based. And then they just need to create the visa sponsorship and the right opportunities for you to move over to UK. It's such an inspiring story for everybody. Now, Nicholas, let's also dive deeper. You said at the beginning, your challenge was your private jobs online, nobody give you opportunity because you're in base in Brazil or maybe your resume wasn't well written, maybe has nothing to do with Brazil, right? So tell us what kind of challenges have you faced when you land your next job as head of AI in an international company? And how did you conquer those challenges? At the very beginning, I was trying to go get into the interviews and then try to have the conversation. And after that, when I asked it, if I had the H1B visa or everything, then, yeah, say the truth, of course, I don't have it. I work remotely, I work for a company remotely. And I spent a lot of time with that, with these interviews. What I changed in my behavior was, I started thinking that maybe one curious case was a recruiter that approached me and then called me on the phone to interview. And instead of interviewing me, he spent the first 15 minutes of the one hour we had together, just selling the company, saying, this company is so good for you, your experiences are great, you're the best person, we need you to start right now. So he spent 15 minutes just selling the company and I was like, uh-huh, uh-huh, okay. In the last minute, he asked it, okay, so just to be sure, are you right now in US? Are you a US citizen? Then I said, no, I am not an US citizen. I am Brazilian citizen. I work for a company that's based in US remotely. And then he went from 100 excitement to zero, like, ah, okay, so I call you back in three weeks. Then that changed some buttons here on my brain and started thinking, well, maybe I just losing time with this, so I need to get into a company or place or talk with people that truly appreciate the intercultural improvement that hiring someone from outside US can bring to the company. Yeah. And then I stopped going into all interviews and then I started saying right away, I'm Brazilian, I don't have the H1B and I don't have citizenship. And then I started after setting everything up as I learned in the PMA course, I was receiving a communication from recruiters two to three times a week. So they sent sending rows directly to me. Then I started just saying, I'm not a US citizen. If this isn't a problem for you, let me know, we can schedule a meeting. So here's my resume, okay. And then I filtered out 90% of propositions, saved a lot of time and went into real opportunities. I found out I had some offers from companies that were interesting in the health tax sector, but not exactly what I wanted. So I just waited a little bit more and when the recruiter from Save the Children approached me and then I learned about the organization I fell in love with Insta, I fell in love with the story and this organization. It's a 104 years old organization with 17,000 people. And before I was in a startup, very fast for Agile with hundreds and a few people. So it's a big change for me. And I'm learning that they really see with good eyes the PM experience. This is awesome. So basically conquers the first challenge in terms of visa, people do not embrace your international experience. Can you also tell us more regarding how you land ahead of AI role? Because the two challenges, there's lots of even Americans, they don't need visa sponsorship. Most of them couldn't land ahead of AI role. So how do you actually climb up to become the head of AI? Especially you're just senior PM, right? So now we become head of AI in a such large international organization, move you to UK, you must be much better than other candidates. Can you tell us more regarding how you made it happen and how to really make you stand out compared with other candidates? Okay, so I always focus a lot on impact. I think there is a big hype around AI always head and people think that displaying ability with algorithms and techniques will make them a good AI PM. But that's not the case. To be a good or a great AI PM, you need to be always focused on the problem. You're trying to solve in the impact you're going to generate with that. All the time question ourself, am I solving a problem or am I looking for a problem to use my tool? So AI should be the last tool you decide to use because of the complexity with data versioning, with software versioning, with how the user is going to interact with it, the UI, UX around the AI is very hard. So you should always try to not use AI, use simpler solutions. Of course, I felt a lot before in my career, I felt a lot in the trap of trying to over complicate things until I learned out that the most important things to generate impact, to drive impact, to move pointers, KPIs. And I was able to focus on that. And in the interviews, when asked about technicalities, I always started reaffirming, reassuring that the most important thing is the pointers were trying to move the impact we want to generate. In the case AI is necessary, then I would work like that. And I think it was important for a big organization like that not to have someone that is blind, blind by the excessive brightness of technology. I think this is the key phrase, don't let yourself blind, get blinded by the excessive brightness of technology. Awesome, awesome. You actually share this with our alumni panel for our private communities inside PMA. Regarding now we have AI, you're selecting the best AI for a manager to join the company. You set up a trick question, right? Can you set up, can you tell us the trick questions? You set up and show them, as an example, what do you mean by, hey, AI is the last solution, but lots of people just wanna dive into AI and show them, I noticed AI and everybody has AI. Everyone's every startup's company, they say, oh, where AI start up, they can raise more funding. Everyone just wanna add AI to it, but sometimes it's not necessary. So show us the example, show us within a private conversation and take the best AI candidate. Yeah, okay, so at that time, I was a lot of the time in the side of the person who hires someone. And even for data scientists, at that time I was hiring a data scientist. I don't go all the way only, I don't look only to their ability with algorithms and their understanding on that. So I had a trick to see if the person is actually thinking or is just repeating things. So one tricky question I came up is, actually it's a three-part question. First I ask, if you had an exam and you had to use to detect, for instance, a nodule in this image, what technique would you use? Then the person answers convolutional neural networks, okay, cool, so this is the best result, great. If you need, then I went to the second question. What, if you need to detect horns in a sound, what technique would you use? Then the person answers, recurrent neural networks. Great, okay, that works. And then I came with the final killing question with 90% of people failed directly. Okay, so now which technique or solution would you use to control the traffic lights in a crossing? Then the person answers, ah, columnar AI. No, man, no, we prove it instantly because it's statistical, it's an statistical tool. It will fail no matter what. 1% of the time, 5% of the time people will die. So you'll never use AI for a critical application like that. And it was very important because I was working in the health tech sector. So you don't have room for mistakes. You need to know when can you be fast, when can you use AI, and when do you need to go slow to be very, very careful with what you're doing. That's the trick, yeah. Very smart question. So basically you're like, all other questions about true AI questions. Last question, you shouldn't use AI for traffic lights control at all, or even flight control. I have some student come up, oh, why don't we use AI to control the flight pattern, like Delta flight, United flight, use AI to control the flight of the airplane. I said, oh, then we're gonna have airplane crashes soon. Sooner or later, or whatever small percentage chance will happen because AI sometimes, as you said, is a mathematical algorithm. Sometimes it may make mistakes, but when it makes a mistake or mission-critical things such as flight, or healthcare, or even traffic lights, it cannot happen at all. It's not horrible. So people just want to step in AI as a keyword to show them they know something, but they will just trick by your smart questions. That's very smart. 90% of people fell on the spot use AI to control traffic lights. Cool, I learned something new. Everybody, if you like this new interview, trick questions, then comment down below and like this video. Let us know what do you think. All right, so Nick? I have a lot of them. You have a lot of them. Comment below. If you want to hear more, we want to make more video about tricks. Massfell, AI interview questions, 95% of them failed. What about you? Awesome. Okay, cool. So now, I'd like to talk more regarding what specific most important shift that pushed you to the next level? For example, as you jumped from just near PM to kind of AI in a large international company, what do you think is the most important shift that pushed you to the next level? I think the most important shift was to, there were two that I can see looking retroactively right now. The first one is that I started to understood my career as a product. I always said I am not a good salesperson. I'm not a good negotiator. I am a good product manager. So this, how to manage a product, I know how to do to look at the KPIs to troubleshoot what's happening, why interview, get feedback. When I just something click in my mind, I think maybe one tip or trick that someone gave me in the community, then I started to understand, oh, okay, so if my career is a product or this journey of getting a better opportunity is a product, I need to understand my funnel. Where am I failing in, or where am I losing the opportunity in the funnel? What's the conversion rates expected from sending resumes to get interviews, to getting a second interview, to getting to the hiring manager, and then getting offers? And I selected KPIs and used the linkage to understand how many people were visiting my profile, how many recruiters a week. So I use these techniques and was able to get to the bottlenecks and then went from two months without any interview to two, three interviews a week. So it were small things. It's not worth that I say it here because for every person it's going to be different. But what is important is to understand where are these real knots, little knots that are preventing you. Another knot that was preventing me from success is cultural traits that we Brazilians have. So it's for us, it's very arrogant to say, I have done this, I have done that. So even for our interviews in Brazil with Brazilians, if the person just starts talking, I did that, I did X, Y and Z. We think, well, what about the others, the team? So we directly understand that this person is not a good team player. And for us Brazilians collaboration is extremely important. So, but for US I learned that it's different. So I needed to say, I did that because I did that and it's not, there's no shame on that. So that's helping me a lot. I think because for a new US company or others, if you just say we did that, they think different. They think that probably you are only one more person in a big team and probably didn't do nothing for real. Yeah, exactly. I like you highlight those culture differences. Actually in the US, you're right, it's opposite. When you're going to interview, you say, we did this. They didn't know what exactly you did. Did you feel like you might be the one who hiding behind, take other people's credit? And US culture is also very big on entrepreneurship and leadership and leader always. I led a team to do this. Yes, a team of 20 people, but I led them to do this. All right, it's more about help you to stand up from the competition, but you're right. I'm Chinese culture as well, we try to be humble. And it doesn't work in other countries, but only work in US cultures is very, very special. I'm glad you brought this up. So people are more aware of the differences between different cultures and also know how to navigate when you jump from countries to countries. So Nicholas, you have started your head of AI position in a new job. So can you tell us, give us high level overview? What does the AIPM do in your day to day? So what an AIPM does is exactly what the APM does, plus understand or be prepared for the additional challenge that can come with AI. These additional challenges are, for instance, now you have to deal with the data. And so the data lifecycle, you need to deal with engineers over complicating things. I can say because I'm an engineer and I over complicated things in the past. So engineers will always want to use the latest technology, but as an AIPM, you need to know enough to ask the questions that simplify the problems that target people to what is the impact that you are trying to drive. And the AIPM must also understand how do you make sure that your developments in AI are going to reach production and generate the impact you want. It's hard because you don't know exactly sometimes the level of data literacy, the knowledge the final user is going to have regarding AI. So these are questions you need to ask yourself. What's the level of explainability that will be needed? For instance, just a label saying you'll have a disease or not, it will suffice if you are building an AI system for a specialist radiologist doctor. Probably yes, but for a technician, you will need to provide the heat maps and the point exactly what's happening. But then comes the crucial question after that. Okay, so your user, understand what the AI is saying. Is he in a position to do any kind of action that generates value for him or for his company or for your company? Maybe not. So this AI prediction, the solution maybe should not exist. So I think this is the hardest part to understand. You start with the problem, but after that you need to ask what action do I need from the user to be triggered with the help of AI, achieve a solution for this problem. I see, so sounds like it's less about hey, let's hire some smart PhDs from MIT to build AI model. And it's not about hey, let's hire some smart engineer from Stanford to train the AI model to collecting data. It's more your understanding to collect data or to make sure the data is clean. But harder question is once you bring it over to the end users, how do they use it? And also is this the right interaction interface for them to use it? And also how exactly they can make decisions using AI, am I right? Yes, and you need to take it to have in mind that there are three levels of AI implementation. The first one is when you use AI to improve an existing feature or empower that. The second level which is deeper is when you build a new feature that wouldn't be possible without AI. And the third level is when you build an entire product that wouldn't be possible with AI. So AI PMs are most needed here in the third. If you are in the first level just improving an already existing feature that already generates value, probably you just need a data scientist or maybe a developer using an off the shelf solution. But the AI PMs really shine when you use AI to produce something that's completely new. An example of that are, for instance, self-driving cars. Sure. Something like that, yeah. Exactly, actually, we've talked several self-driving car live cases instead of PM accelerator talking about how AI is being happy to use in self-driving cars. Actually, I happened to build the very similar technology called machine vision in my first PM job as AI product manager and also the same technology also used for self-driving cars because the original one was more for smart cities to reduce car crashes, but the same thing being used by self-driving cars. You're absolutely right. This is a foundation for self-driving cars even exist running on top of AI, but some applications, if the marginal improvement, you may not need the AI product manager, maybe just existing product manager hire some data engineers as something new to it, such as something called Notion AI Fund, very funny. Notion is just no taking tool there. It's a very, very good organization tool. They just call themselves Notion AI and then have some kind of AI in it and become an AI company, of course, good for them. So it helps them to go IPO, increase the stock price, but it's not the kind of what you described, the very necessary true value of AI product manager to create for any company. Thank you for your insight with that. This is awesome. Yeah. There is one other insight that I'd like to share. I think I could also summarize the whole of an AI PM as someone that can understand the data strategy, match this with the product strategy and generate a business mode, so MOAT. So a sustained advantage for the company over time and leverages the power of the data to the core of this mode. So you need to build cycles like this. The more data you have, the better is your product. The better is your product, more users you have, more data you have, more or better data as well. So if you can build the cycle in which the quality of your product will always improve with the amount of data you have, so this data will be private, it will be generated within, it's monitored that just generate this data, part of it is generated within the company product. So the company will have something that others don't have. So technology by itself is never a mode, is never a sustained advantage, but this technology together with the data align it with the product strategy, then you'll have something that truly shines. It's an advantage like the network effect. So you have the network effect with providers and users. You put data on that and then it's the same thing, looking through, solve through another list. Awesome, I like how you teach other people using very simple concept to understand the cycles and continues to grow in the bigger group of stronger AI ecosystem. This is amazing. Nicholas, let me ask you. Nicholas, now as head of generative AI, what kind of advice do you have for people who want to break into product management or who want to become an AI PM? First of all, what I learned and you will probably learn as well, should learn is that if you are someone that is a generalist and wants to build things and make sure that these things achieve people that get in production, generate value for the person, for your company. If you're someone that has this mindset, chances are that you are probably already a product manager, you just don't know that yet. I was really lost in my career for many years because I wanted to do everything, but there was no specific role for this do everything person that is the glue between teams and outside teams. And then I learned about the PM role and then I thought, oh, okay, that's it. So there is a role that I can follow people, I can get courses on that. The first thing is to acknowledge if you have these aspirations and if you like to work like that, then get some formal training first. I started with a professor, Alex Cole, from Virginia Darden Business School, which helped me to organize my thoughts a lot and learn a lot of new tools. And then a few years later, with your help, Dr. Nestli, I was able to understand how to wrap this, all of this and leverage this for getting better good opportunities in the market. First of all, understand that. If this is your career PM, then get some formal training and okay, now becoming an AI team, which is more specific. I would spend some time or invest some time understand learning about the off the shelf tools first. What are the tools, the things you can use off the shelf? For instance, detecting things on images, detecting sounds in OCR, a lot of these things you have a lot of solutions within Google, Azure, AWS, you won't need to spend too much to use that do some proof of concept, maybe if you know how to code. And it's important, I think, to look, I think that data science won't be looking. How does the pricing strategy or rules work for the solutions? I found out a lot of solutions that look very cheap at the beginning, and then when you put in production that the cost scale very fast, the same things happening with generative AI already. I saw some companies with some pricing strategies. For instance, they charge very cheap for input tokens, but they are very expensive for output tokens. And they are a company that deliver proposals, for instance, so they generate a lot of tax. So it gets expensive very, very fast. So understand pricing. One experience I had was that at some point our company was stuck because the AI models were too expensive. We used it, the best tools available, that the big tax that was accelerating us suggested and the great work it, but then we were not able to launch more models because it was too expensive. We were with one million exams a year with a cost of $200 per model per month. Then we migrated to another tool that doesn't make the AI model online all the time, just scales up and scales down. And they were able to reduce this from $200 to less than $2. So because our peak of usage was very, very narrow. So I think if you start to develop these skills, understand the pricing of the off-shelf solutions, organize that on our mind, ask the question. I think as a PM you are already transitioning to an AI PM role. This is the very beginning. And then you get deeper with your engineers, you learn from them and then become even better. Another important thing is to start to get to understand, but I don't think this is trainable or you can study. You'll get with experience, the know-how, the feeling, when to pursue an in-house solution, when to hire a solution from a provider, a small startup provider, and when to wait for a big tech to launch the same solution. I see. Basically, there's a steps of tiers people need to do, right? You must understand the basic product management be a great product manager first, and then you can make AI specific decision in terms of pricing, existing tools, which wants to use basically usability of customers. And then you figure out how would you try to like in-house and outsource what kind of more high-level strategy decisions you need to make as AI PM. Awesome, this is great, cool. Hey guys, you know that Nicholas is going to film a second video talk about index regarding general AI for everyone training for everybody. So make sure to like and subscribe this video and also check out our next video right here that is the general AI training for everybody. So you can also become AI product manager. You're going to dive into all those three steps we just described here. Looking forward to seeing you in our next video. And thank you so much, Nicholas, for joining us today. If people want to find you, where can they find you? Okay, so LinkedIn, I don't have any other, I already have Medium as well. So sometimes I post these thoughts, these insights I have there. So Medium and LinkedIn, I think is the most. Awesome, yeah, reach out to him on Medium and LinkedIn. I'm going to share all the links in the description of this video as well. Thank you so much for joining us, Nicholas. We're going to see you in our next video right here. We've got an in-depth training of generative AI for everybody. This is darknessV from PMXCenter.io. We'll see you in my next video right here.