I’m so happy to have had the opportunity to interview Tina Nguyen for this podcast. In this episode, Tina shares how she came from humble beginnings in Vietnam to go to college, receiving multiple master's degrees, an MBA, and now is an ML/AI Portfolio Manager at U.S. Bank in their Data Science department. Tina calls the women in her life a major strength in helping set her work ethic to arrive where she is at today.
She is also on the executive board of directors at Northpoint Health and Wellness Center and a Woman of the Year Legends Award Winner from the Leukemia and Lymphoma society.
In an extremely inspirational interview, you'll learn that Tina is a real go-getter and I fully believe that everything she sets her sights on, she will achieve.
If you are interested in learning about how AI is being applied across multiple industries, be sure to join us at a future Applied AI Monthly meetup and help support us so we can make future Emerging Technologies North non-profit events!
Resources and Topics Mentioned in this Episode
Tina Nguyen 0:00
My grandmother and my mom always said, You are very, very lucky. Number one to be alive. Number two, to not have any restriction put on you because other women in Vietnam, when I travel all over the world to kind of talk about women rights to a lot of girls were surprised like, Oh my god, you know, you went to the US by yourself. When you were 14, you had your MBA and your second master by yourself, you travel to all these continents by yourself. And I'm like, Yes, and you can do the same thing.
AI Announcer 0:32
Welcome to the conversations on applied AI podcast where Justin Grammens and the team at emerging technologies North talk with experts in the fields of artificial intelligence and deep learning. In each episode, we cut through the hype and dive into how these technologies are being applied to real world problems today. We hope that you find this episode educational and applicable to your industry and connect with us to learn more about our organization at applied ai.mn. Enjoy.
Justin Grammens 1:02
Welcome to the conversations on applied AI podcast. Today we're speaking with Tina when Tina is an ml AI Portfolio Manager at US Bank, where she manages the Enterprise's artificial intelligence and machine learning solution portfolios, and helps build the vision strategy and roadmap of AI ml and analytics across the enterprise to drive impact for all business lines. She was also on the executive board of directors for NorthPoint Health and Wellness Center and a woman of the year legends award winner from the Leukemia and Lymphoma Society. She's also a fellow graduate of Oxford college like myself, so go augis. Yeah, really awesome. And I'm really happy to have you here. Tina, welcome.
Tina Nguyen 1:39
Thank you so much for having me, Justin.
Justin Grammens 1:41
So you know, I gave a little bit of background in terms of like where you're at today, I don't know, you want to fill in some of the some of the dots, I guess, with regards to how you got to where you're at from school? And also, I guess, how are you seeing AI being applied in your current work?
Tina Nguyen 1:55
Yes. It's been a great journey, actually. So I started at Oxburgh bass experience of my life really, truly, I was major in international business and math. And actually, in my head, I was imagining, you know, I would be traveling all over the world, and using math to solve humanitarian issues. You know, coming from Vietnam, we have a lot of you talk about right now, at this moment. This time we'll be living in. You talk about women rights, you talk about people living in poor conditions, and not being representative of the entire populations. You talk about children's rights and sexual harassment and all that kind of thing. And that's what I'm passionate about. So being in the UN Oxburgh really kind of pushed me forward using what I'm supposed to be learning in mathematics and business. To put it together. Somehow, I didn't know how this would all go together. But I knew that was the combination I wanted, partly because I just love math so much. And that's how I came to the states is because I was a champion in my class in Vietnam for English literature, or mathematics and English. Oh, wow. Just the language. Yeah. So I got chosen and I came to the states when I was 14. And from there kind of Americans raised me strangers raised me in this country. So went to Augsburg when I was 16. And study math and business, like I said, and started working at US Bank right after Oxburgh. And from there, just a book personal banker started studying everything about the banking world, working with customers facing with customers, and I learned about the personal banking side of things. And then I really was interested in the business, the commercial side, so I move over to become a business banker did very well in sales. But after that, you know, I asked a ton of questions within the platform that we use, there's a lot of lead list, you can call, like, hey, call this person for this product, call this person for this product and whatnot. And I always ask the question, why? Why do we choose certain customers for certain products or certain messages, and my manager at that point of the brand, just, you know, he's like, there's a huge algorithm and a huge team, I can understand I can understand it either. But um, you should come and ask them. So So I basically hunted down the Strategy Group from the business side, the manager of that group, Mike, can you tell me how that really came into play? Like, what was the strategy behind it? What were the data that you collected? Why are these customer selected for this particular product? And then he's like, Well, you know, you can join my team and see how it goes. So that's how I moved from sales into corporate world. And I started with a small business banking Strategy Group kind of learn how they collect data, how to analyze and that's really how I started the analytical path. So I looked at all of the information they had and the data was very scattered at that time, small business was still very much that was Like back in 2010, where small business was the focus, but a lot of people don't know what to do with that data, right? A customer is easy. You just have a social security number, your name, your address, and all that kind of thing for business, you have a holding company that holds multiple children, businesses, or you can have Tina LLC, and then on the side as a Tina, selling, I don't know, Girl Scout cookies, and you have to have that kind of business on the side, but you cannot treat them the same way. Sure. So I helped design a business owner view using data collection for US Bank, and then from there created basically a holistic view of the business owner behind all of this businesses. Because when I looked at the strategy at the beginning, you have a lot of silos treating different types of business separately, right, based on their sales, their type of businesses, and all that kind of thing, which makes sense. Any banks do that. Any financial industries, companies do that. But I'm always about the human behind that, like, okay, so you calling all these companies, but there's a connection link, there's a person underneath it, how do we get to that person so that when you call one person, you don't have five different bankers calling five different businesses that they have, the CEO is going to be maintained, and connected and deepened relationship with us once we have that. So that's how I really learned to dig into the data, you know, data wrangling, cleaning up things matching, fuzzy matching, and all that kind of thing. So then that's when I started learning more about predictive modeling. Because like, Okay, once you figure out the customer, how do you figure out who's the one to go after? Right? Because so say, I'm just throwing numbers, we have millions and millions of marketing dollars out there? Well, should we spend that on everybody? Of course not. So we have to choose a population. So how do you segment those customers out? So I asked the question, and like, we don't know, you know, there's a bunch of marketing mix models out there. There's something called mmm marketing mix modeling, there's MTA multi touch attribution, but really is the last touch. So basically thinking about the customer path. And then all the touch points are, you know, you you wake up in the morning, you turn on the music is that ads on there, and then you brush your teeth, you go down to get to make your coffee, and then you drive to your work. And then there's billboards out there. So we touched another customer, you pass a US bank branch, how do you measure that entire thing? And then somehow that customer convert, come and buy a product with us? How do you know which touch point? is the most important touch point? Or is it the last touch? You know, is it that last thing that we did? Or is it that three years or two months before or whatever, that we planted the seed that kind of build it together?
Justin Grammens 7:46
So you you were looking at it mainly from a marketing standpoint, right? using data to market to customers at the right time at the right place to be in most effective.
Tina Nguyen 7:54
Exactly. And that is built by modeling. So that's when I started being very interested in machine learning and predictive modeling. And that was when the AI thing was not that big. Anything in terms of mathematics and whatnot is like actuary works, right? So I'm like, I do not want to be an actuary. It's just not my personality, I cannot do that. So then I finally found a modeling team within us bank, it was a very small team of like five people, led by a woman, and all four were males.
Justin Grammens 8:29
What's the timeline? around? Like, what year was this?
Tina Nguyen 8:31
Let's see, I think it was 2012, or something I forgot it was on my resume. It was been a while.
Justin Grammens 8:40
I'm just always sort of fascinated with regards to like, I guess, people think of this AI and this modeling. It's like something that just happened a couple of years ago, right. And the more I talked to people, people were doing this, you know, in 2010 1112, that like yourself, and and, and like, like others, they were sort of the people that were pioneering this sort of machine learning space before, it was really called machine learning and deep learning and all this stuff. Obviously, AI has gone through a series of you know, it can be used in multiple ways, and it's gone up and down and stuff like that, but it's safe to say that, you know, you and others were were looking at it from this perspective, like a decade ago, right?
Tina Nguyen 9:15
Oh, yes. Yeah. And I mean, if you came out of math, right, applied math or discrete math, you learn all of this all the time you learn about graph theories, you learn about the measurement, you know, like, measuring, I don't know, the cosine degrees and all of that stuff. And now people are doing you like, oh,
Justin Grammens 9:35
oh, yeah. Yeah, like differential equations, and basically integrals and all that type of stuff. Right. Cool. And I majored in Applied Math at Augsburg. And so you're doing this stuff and you're like, not really understanding how you're going to apply it. And then, like you, the only real job that I had coming out of school was either an actuary or go for my Masters, right? And now it feels like there's just so many other opportunity But you apply all that stuff, you know, every day, especially in like regression models and stuff like that you're you're really trying to find the best path, basically. Right? So it's, it's cool to see how it all can actually come back to you. And now math is sexy, I guess in some ways they say, right,
Tina Nguyen 10:15
we are the cooler guys now. Then I basically harassed the manager, the former manager for the modeling team actually moved on, she got promoted. So one of the colleagues, one of her direct got promoted punkish. And I just kind of harassed him for I think, a whole six months and Mike punkish, give me a chance. Like, I know, I don't have the background of that. I haven't done any predictive modeling. But please give me a chance. And he's like, well, Tina, you know, you a lot of our team members are like PhDs in mathematics and all that stuff. Do you have a master's degree, that's when I just finished my master's degree at Hamlin University for just an MBA, not just an MBA for an MBA, and then I was doing my second master at U of M for system engineering. So I want it to be a petroleum engineer for some reason, but then that completely did not work out. And so he's like, you know, but we do have a need that we start to see. So this is the key moment of my life. And I did not know that. So he said, we built models to predict all of this cool things. Nobody knows how to apply that. Or when they apply the model, they apply it incorrectly. So often, the reportings come out differently, the population of the customer that we originally built for completely change. And then they come back and they sell your model sucks, you know, he's like, no, the model stack rank exactly what is supposed to stack rank, its function, we validated every month, and we make sure it works. So there is a disconnect between the data scientist and the business and how it's applied. He's like, oh, help me figure that out. And I'll give you that chance. Oh, my God, I can take that challenge. And I did. So I came in. At that time, when the whole org change happened. It was just punkish. I was his employee, number one. And then he hired two more data scientists after me. But we built from ground zero at that point, the modeling team that US Bank now has the enterprise data science team. And yeah, and that's how it all started.
Justin Grammens 12:21
Awesome. Well, that's a great story. Now, you know, as you were sort of talking, I was sort of like thinking through, you seem to be a person that's pretty well versed, I guess, right? You You're thinking technically, with a mathematics background, you enjoy being a personal banker, I guess, or you enjoyed the human side of all of that, I guess, and then sort of applying all of this data to actually then change people's lives or help them out. Right. Is that kind of a fair thing to say?
Tina Nguyen 12:45
Yes, thank you,
Justin Grammens 12:47
I guess what are some few words that would describe yourself? And in some ways, I guess, what do you feel that maybe are your greatest strengths and weaknesses? As you've kind of like looking back on what you've done over the past 10 years? Plus,
Tina Nguyen 12:57
yeah, so my shirt says, social justice warrior. My husband got it for me as as our wedding anniversary present. So my entire life as I you know, started is I come from a country with a lot of conflict. But at the same time, education has always been the number one you know, it the Asian stereotype, right? And it's funny because they like, Oh, you must be good at math. I'm like, you know, I'm very good at math. So my, my greatest strength, it has to be the woman in my life. My grandmother raised me as a single woman, while my mom as a single mom went out there and make money to raise two, three children. So grandpa was killed when my mom was four in the Vietnam War. So Grandma, basically, oh, my God, her story is insane. And she raised my mother adopted my uncle and adoption at that point was not even like a thing in Vietnam. And they said, why would you bring a child, you know, and she saw my uncle on the rice field being fed by mice, because he was abandoned. So she just took him home. And she's like, Oh, right. That is my son. And so then my mom was abandoned by my father when I was four. And my mom then took on two three children, she divorced him an alcoholic at that point. First divorce ever happened and dosti I know it sounds very weird that I'm so prideful over it, but my mother was, you know, she's like, yeah, you know what, I want the best for my three children and I cannot have that happen. So my mom and my grandmother raised the three of us and First thing first she's like, okay, you have to be top three in your school. It was supposed to be top one actually. And my sister is the the rule follower. So she's always number one in every single class. But I'm like, you know, Mom, I, I cannot be top three. I have to play two and I have friends. I have things I have to do. Can I be top And she's like, Okay, fine. Top three is
Unknown Speaker 15:02
Tina Nguyen 15:04
Yes, still good enough. So my entire life we study English I believe when I, we had a tutor, English tutor come to our house full time when I was five. So that's why my, our English in our home is very much fluent. And math and science always, always on top every year, I think the moment I can calculate and do things, I think when I was little, I was doing multiplication in first grade. And my mom was like, okay, for all the classes that you are taking, I want top in mathematics, biology, physics, and English. That was her requirement. And we were just like, okay, that's, that's it. So that was my strength is the woman in my family the value that instill in us this, you know, my grandmother, and my mom always said, You are very, very lucky. Number one to be alive. Number two, to not have any restriction put on you because other women in Vietnam, when I travel all over the world to kind of talk about women rights to a lot of girls were surprised like, Oh, my God, you know, you went to the US by yourself. When you were 14, you had your you pet, you know, your MBA and your second master by yourself. You travel to all these continents by yourself? And I'm like, Yes. And you can do the same thing. So that is my strength. I think my weakness, I must say, I'm too persistent.
Unknown Speaker 16:31
Point of annoying.
Justin Grammens 16:34
Well, got you to where you were with regards to your current job. It seemed like right, you. Give me a chance. Give me a chance.
Tina Nguyen 16:40
Yeah, so once I have my eye on you, I think you're gonna have to say yes, otherwise, you're screwed.
Justin Grammens 16:47
That's great. That's great. Well, at least I get I don't know if that's a weakness. But but Sure, I can see if you, I don't know, it could it could be taken the wrong way. But But again, it's good that you know about it, and you know, you're able to handle it and manage it. But it can definitely get you far in life. I think being persistent. Sounds like you're a real go getter for sure. What's I guess? A couple things, bring it back to AI. two things. One is sort of like, how would you define artificial intelligence? So, you know, you're, you've got all this data, right? And your guys are working and collecting it? What makes it be intelligent, I guess, how does it move into machine learning? They may guess those are the tools that you use, but you kind of have a definition with regards to maybe what you feel AI is?
Tina Nguyen 17:29
Yes. So I think a lot of time people want AI, quotation AI here. And now people knows, like you said, the definition of it, and the implication of it, right? So when you think about AI, artificial intelligence. So you, you watch a child read or do something. And in your head, you're like, how does that child perceive things? Right? So AI would be doing a human activities, anything driving, talking, communicating, painting, even, but using the image, emotion and the human side of it to complete the task. Now, machine learning deep learning is different, though, right? It's a subset of AI. I don't believe we currently have any AI solutions out there. But we would say get successful. There's no algorithm out there at this one that I know of, and maybe I'm super wrong, and I'm behind the curve here. Let's take chatbot for example. You know, when you chat now, a lot of design for chat boxes, like they actually intentionally misspell something. So that the other side human interaction when we type to chatbot, we think, oh, that's really a human behind it, but it's not. It's still designed that way. But if I say, you know, somebody, say, Oh, that's funny. There is no such AI robot out there that can catch that tones and said, Okay, that's actually really bad. When she said, That's funny. It's actually not funny. So please, escalate versus Okay. Well, thank you hang up now, you know,
Justin Grammens 19:04
right here, because he sort of taking the word at its face value, not really how its presented.
Tina Nguyen 19:10
Exactly. So I think I was just listening to your podcast with somebody, Jake. Yeah. Jake, and his talk about you know, we just give it to God and hope it works. So I don't think there's AI yet. And there's the fear of a robot taking away jobs human, I want to say to everybody, that is the stupidest thing I've ever heard in my entire life. It's the same thing as saying immigrants are coming to the US taking away jobs. I'm an immigrant here. I'm working in this field. And I know for a fact my job will never ever at this point, be replaced by a robot because I have emotion because I have relationship. I have connections with people that a robot will not ever be able to replace. Therefore, the human connection to human behind every algorithm will never be replaced. Whatever you do to automate, you automate documents contract under NLP works. So you're talking about use cases, right? Some of the stuff that we doing is automation, for all the manual work that we're doing a lot, because like, Oh my god, you replacing my jobs, right? We can say, No, it's not replacing your jobs is really making your job more efficient, instead of you looking through hundreds of documents and pages, to extract these type of what we call entities, right for entity recognition here, like the name, the addresses and all that stuff, just for an audit purpose, you can actually now look at a screen and the algorithm picked it out for you. And what you can do is the enhancements of that you can actually use your brain versus having to keep replicating doing the same boring job again. So no, we're not going to replace you. However, if you decided you don't want to learn anymore, then unfortunately, that would be a problem for you.
Justin Grammens 21:03
Right? Right. You know, as you were speaking, I was thinking about, there's a book by Andrew Yang called the war on normal people. And he was a presidential candidate, and I'm not sure if you're familiar at all with Yeah, it's a very interesting book. I mean, he a lot of things that he talks about is this is this basic income that people should be getting, like, $10,000, a year, something like that. But all of that stuff aside, he see lays out a really, really interesting, basically a hypothesis here that going forward, a lot of these jobs are going to be automated, right? And so there are people are going to be automated, like just out of jobs. And the hardest thing is is can they make the leap? Can Can they move into a career that is more creative than that has all of these things you talked about? You know, you're basically talking about emotional relationships, you know, connections, those are things that a lot of humans have, but they also need to be creative. I think that's the biggest problem that I've seen with machine learning and AI, there's none of that creativity going on. And so you know, how can we incentivize people help educate them help at least get them to a minimum basic level, because if somebody is in poverty, and all of a sudden their job has been taken, because of automation, that is that that's not good. And if you can bring them up to a certain level, there is a chance for them to succeed. But yeah, I completely agree with you. There's a lot of jobs, everything from you know, flying airplanes to farming, right. So farmers don't want to do the the minutia, they would love to hop in a tractor and let the tractor take over the hard work. Same thing with call centers, like who really wants to be in a call center these days and take all these calls, why not have a chatbot? handle it for me, right? So all that stuff can be automated away? But I think there is the biggest question that people have then is then so what do I do now? Right? are we all going to become programmers? are we all going to become engineers? Do we all need to understand technology? Or what's the path forward? Do you think on on some of those roles and jobs? Or do they do they just change and morph?
Tina Nguyen 22:52
Oh, yes, I believe evolution is a part of who we are. Okay. Also, you know, I'm bringing it back to immigration, partly because I'm an immigrant immigrant myself, when you think about all the immigrants coming to this country or anywhere right now, right? You come here, and you have all just education back in your country. But they are not honored in the United States how to use restart, and yet immigrants restart. And you see how immigrants community here, flourish, also. So if you look at that, that should be where your encouragement and hope comes from. I know friends of mine, who were doctors and lawyers in their country, and coming to the US have to go back to school, retrain everything, we do everything, new language, new culture, new everything. And yet, now they are doing exactly what they want it to do. And so I think that argument that there's nothing else out there, that's just this just unsound. And then I believe every human is creative in our own way. So talk about AI right now, right? So there's a lot of tech than engineers or any ones who are great in mathematics, and biology and physics and all that stuff, right? Go straight to machine learning and AI, I think that's the need that we have in terms of technical stuff. But at the same time, my role is very unique also, but it is much needed for the AI community, is the fact that we can see the universe of stars, and we can make the connection. That's where I stand where I come from is, you know, data is just data, you need to be able to make connections between the data for you to drive a strategy, right? You don't have to be technical. You just have to see the patterns, you need to be able to see the future or you just look at a mess. And you can be like, you know, I can organize this into different clusters of things and it's just treated that way or you have to be curious. That's the only requirement I believe is creativity. If you if you don't think you're creative, that's the But you need to talk to somebody else who really tell you you are creative. But also curious, if you are a person who asks a ton of questions, you are in the right place, and you will make a huge difference because that's exactly what I did is exactly what every single data scientist in my team does, is they are non stop question people.
Justin Grammens 25:20
That's great. Tina. Yeah, for sure. curiosity. It's it's something that I'm trying to instill in my kids. And it's something I believe, as well, as new technologies come out, you got to get on the bandwagon and just start asking how they can guide and how they can be used. And pretty soon, yeah, you find yourself sort of in the midst of it. And all these opportunities sort of can open up for you, as long as you remain curious and continually trying to learn. You mentioned what you do in your in your job like, yes, like, what is a day in the life for for somebody in your position?
Tina Nguyen 25:51
Yes, it's super fun. So if anybody is interested, please contact me. We are in the midst of of planning, neural IPS. So you know about nips, right? The conference icml and newer IPS, US paying for the first time ever will be a platinum sponsor at the nips conference 2020 in December, and my team are the team that's making this happen. So we will have two panels a day in the life in the data scientist at US Bank. That's one of the panel discussion. And then another panel is really how do you use all of this complicated algorithm and the text behind that and apply it in real world, especially in the financial world? So if you're interested, I can send you those information and the date and time for that.
Justin Grammens 26:40
Yeah, I have like liner notes when I publish this. So I'll be sure to include all that information for sure.
Tina Nguyen 26:45
Awesome. So we planning nibs out? So the branding piece of it I you know, we have to talk about use cases. Okay, so how vague can we be, but how specific can we be because the target audience are all PhDs researchers in the world, right? doing all of this techie, and machine learning and AI algorithm development. So, you know, you talked about one interesting use case we're working on is ATM skimming. And that's this video analytics kind of work that we're doing. So a lot of people and you know, when you go to, to an ATM or a gas station, you put your cards in there. That's what the crooks have really design, like new things that they can insert into the ATM machine, you don't even know right, you know, another layer that they put in there. So when you put your card in there, they reading that card, right, and then some of them went even higher, they actually install a camera on top of that, too. So when you put your PIN number on there, they record that pin number. And the whole entire transactions. If you Google ATM skimmer on YouTube, it takes about less than 90 seconds, it comes in, put it in, and then they leave it out for a week or so think about gas station and ATM, people put in their carts, hundreds and hundreds of that information is being stored. And then the guy came back with you know, and get all that information. And then he comes out. And now he has hundreds of people data out there to steal and monies and everything. So that's one of the technology that we are we worked on, okay suddenly improved out. So now we can have our algorithm running in the background and catch the moment an ATM skimmers inserted.
Justin Grammens 28:31
Yeah, so when they want as soon as they start messing with the ATM, you're able to sense it within
Tina Nguyen 28:35
that 90 seconds. Yeah, we sensed it, we sent an alert out. And then we stopped all transaction for customers to protect them. So again, the human side of things, we do cool, you know, video things and image recognitions, and all that kind of thing in the background. But really, my team is a team of people who said, Why do we do what we do? Sure, sure. And who are we helping out here? You know, when you talk about ethical AI and all that stuff, that's always in the heart of every conversation, whenever we sit down, or whenever there's a use case come to me and said, Tina, we want to do this, you know, our team sit together and said, Okay, how are we going to do this? What are the algorithm design? Where are the data, you know, the typical questions, but should we do this? And if we were to do this, how are we going to maintain this and making sure that the application of it is ethical, non bias and at the same time make money for a company?
Justin Grammens 29:30
That's fabulous? Yeah, no, I mean, why Why? Why apply technology if it's not actually going to improve experience for your customers, for the greater good, I guess of society. So yes, be great. Very cool. Well, that's, that's super exciting that US Bank is gonna be a part of that conference. I also I think I saw you're gonna be a part of Def Jam thing as well, too. Right. Is that that's coming up soon.
Tina Nguyen 29:50
Yes, yes. And I just recruited one of our top senior data scientists on their symmetry so he will be my code jammer. So please join and support us. And we treat his incredible new addition to our team. And he and I are starting AI for good kind of group within our team to to do more of that. And what else can we use? What we already established to help people, right. So, you know, my boss worked on a project a while a long time ago to help with sex trafficking rings using AI, you know, and alert the FBI and all those people for activities that we think could be related to that.
Justin Grammens 30:31
Excellent, very cool. I mean, when you were mentioning some of these other applications, where do you think AI falls down? I guess, is there certain places where you've seen and said, Okay, this would be a great thing for us to solve. But for one reason or another, maybe what are those reasons? It just, it isn't a good fit? Or it's not? It's not able to be done?
Tina Nguyen 30:49
Oh, tricky questions? Because this is?
Justin Grammens 30:54
Well, I mean, sometimes I guess, to answer my own question a little bit, you just can't get the data. Right. So you're talking about sex trafficking, I again, I don't know, the data that you're trying to get, whether it's, you know, imagery, or, or GPS location, or, you know, whatever it is, machine learning really works well and has a lot of data to churn through. And if you don't have enough data, it's really not gonna solve your problems, right? Is that fair to say?
Tina Nguyen 31:17
Mm hmm. That is true. So a lot of time we struggle with the data. So you have to really play into the synthetic data field, right? So you have to create these kind of pseudo numbers and whatnot. So our teams is very much into that work as a lot of time and you talk about bias. A lot of our populations, maybe we just it's so new, we don't have that information yet. Right. So you, you have to be creative. You have to be curious and including all that information in there. I think one thing is video recognitions. Facial recognition is one thing that we always whenever we worked on use cases like that, you know, how do we make sure that we maintain the ethical standard, for example, the GDPR rules, right? In Europe and California privacy acts and whatnot. So there are multiple ways that you can apply and not storing images of faces of people. Actually, we worked on a use case to redact that Oh, really? Okay. Yes, of people. Yep. So using computer vision, to kind of blur out all the customers that we need to protect to. So that's one way to use it for good, not for bad, right. And then you also can embed the facial vectors and whatnot, instead of store faces, you can just create numbers out of them and embed that into your algorithm. So there are multiple ways to do that.
Justin Grammens 32:39
Very cool. Yeah. I had thought that on Google Maps or Google streetview, I think you could submit something to them, whether it's your address, or if you were picked up in a picture from their cars driving around, you could actually ask to be blurred out. I remember reading that.
Tina Nguyen 32:54
Yes, that is the rule. And is awesome. So I we support that greatly. And we built algorithms, and we built technology to support that part. So yes. And I know for a fact in Europe, if a European customers in the US even request that they can have their information redacted.
Justin Grammens 33:14
Yeah, for sure. I guess as people are getting into this field, you know, you talked about your trajectory and your path. Are there any classes you see people on your team taking or any books, you mentioned a conference here too, as well, that you might suggest people go to now what are you seeing as interesting things that people should should be doing to sort of educate themselves in this space?
Tina Nguyen 33:33
Yes, for sure. So for the product management, Ai, Product Management and portfolio management on my side, really do go into these conference, right? So icml and neuro IPS are the two top conferences in the world. They have their papers available for you to read. Actually, we have a book club, in our team that we read papers.
Justin Grammens 33:57
That sounds a little dry.
Tina Nguyen 34:00
That's the Sumi tree is the leader for that book club. Actually, that was his idea. So they just every week is like, okay, here's a paper on CNN, go and read it. And then let's talk about that, right. And so that's one thing for my cycle, Product Management, learn about product management. So there's a lot of courses on Udemy, or something like that you can take Coursera for cheap about how to connect the tech world to the business world. Conflict Management is a huge training, I would highly, highly recommend because there's a lot of conflict between business world and the data science world. And you worked in that, you know, a lot of times our clients come in and like whoa, you know, I want this done now, and I want it done in a massive scale and blah, blah, blah, and I want it to be 100% accurate, and we like within you can't build a model. That would be reality. There's no such thing as 99% accuracy, because that would be overfitting and that would be a terrible model to apply. So really learn That right? I am actually reading a book recommended by a friend of mine called one dimensional men. I don't know if you heard of that. It's very old, actually. But it's very relevant to us. It's by Herbert Marcuse. But this book is a reference in a lot of movies, Fight Club, actually, they took a lot of that from that book, basically talking about how the society as a whole because of technology moving so fast, automating things, that we are a part of that machine, unless you make a conscious decision to step out of it, and fight for your individualistic characters and features. It's really hard not to you just like, yeah, you know, I'm graduating from college, I'm getting a house, a car and all that stuff. All sudden, you get roped into this whole thing, as a product manager in the AI world, you need to be able to see where the mass is going, and where you know, you need to stop that or you can create innovative ways for it. So I highly recommend this book kind of dry. But it's great to realize how AI and technology impact the world in a bad or good way.
Justin Grammens 36:14
Yeah, I'll be I'll be sure to include a link to that as well. So Tina, it's been great. It's been great discussion, how do people reach out and connect with you? Is LinkedIn the best place? Do you have other places?
Tina Nguyen 36:25
Yes, LinkedIn is the best place right now. So please feel free to contact me if you want to learn more about AI, Product Management, portfolio management, or if you want to become a data scientist, but you don't know how to make the jump I have a few people have talked to me from academia and say, Hey, you know, I have a lot of research in this. But how do I make the jump from the business side? I've been kind of helping them out with that, too.
Justin Grammens 36:48
That's so awesome that you're giving back. Because there's a lot of people that come out of school, I would say for me, you know, I've been out of school since 96, I guess. So it's been 25 or 24 years or so now. And when I kind of came out of school back then there, there wasn't any internships or anything like that. It was just kind of like you got your degree. And then you're kind of left not to say that the school didn't try. But it's just, you know, the Times have changed. But the more help you can get, from people like yourselves mentors that are in the field, they can open up a whole bunch of new doors and make you think a little bit more about it. So yeah, I'll definitely encourage people to reach out and talk with you. Because your story is really phenomenal regards what you've been able to accomplish.
Tina Nguyen 37:26
Justin Grammens 37:27
Are there any other topics I guess, or projects or anything that you find interesting that you would like want to share? Maybe that you're working on right now, specifically, we've covered a lot, so don't don't feel like you have to but I always kind of asked that at the end? Is there anything that I maybe I didn't touch on that you want it to?
Tina Nguyen 37:41
Um, so I think I did a few talks recently. But you know, when you think about the AI, talent pool, a lot of them are from all over the world. Right? So I am a huge advocate for diversity within our world. I am a multicultural woman. So I think I got a box checked, you know, and then husband always said you got the third box checked? Is you vertically challenged, I'm very short. So you know, you have to have a team of a diverse group of people, otherwise, your model will always be biased, unfortunately. Right? So true. I think as you are looking for jobs, find out your unique identity, find out where in that team, that company that you can be another set of eyes, I think that would be your strength, I would highly recommend people to say, Okay, why am I unique, and everybody is unique in their way. But from the AI world right now is you know, we struggle with a lot of bias. And we struggle with, you know, people not seeing what it is when they building a model from the training data set to the test data sets, and all sudden, when you validate it out, even at the validation point, they don't see that until it come out. And then oh, okay, we discriminating against women in resume. Oh, you know, the Google Chat thing would just start saying racial slurs and all that kind of thing. How do we not learn that? Yeah. So I would say, you know, hire as many people from different backgrounds as possible. And when you are in a team of data scientists look around. Who do you need to bring in in order for your team to make the best algorithm out there that represent the population of the world?
Justin Grammens 39:29
Very nice, Tina, well said, Well, great. I appreciate you being on the conversations on applied AI podcast and wish you the best. Thank you for sharing all your knowledge with our community and look forward to keeping in touch with you and moving forward.
Tina Nguyen 39:41
Thank you so much. It's been an honor. Bye bye.
AI Announcer 39:46
You've listened to another episode of the conversations on applied AI podcast. We hope you are eager to learn more about applying artificial intelligence and deep learning within your organization. You can visit us at applied ai.mn To keep up to date on our events and connect with our amazing community, please don't hesitate to reach out to Justin at applied ai.mn if you are interested in participating in a future episode Thank you for listening
Transcribed by https://otter.ai