Season 4, Ep. 44 – Founder to Founder: Sabber Ahamed, Tech Lead of xoolooloo

(THE FRONTIER THEME PLAYS)

Bill, via previous recording (00:05):

Welcome to another Founder to Founder interview from Gun.io, your source for hiring world class tech talent. Today, Gun.io’s CEO and co-founder, Teja Yenamandra, sits down with Sabber Ahamed, founder and tech lead of Xoolooloo, a service that helps connect you to other people with similar interests in your local area. Okay, here’s Teja. (THE FRONTIER THEME ENDS)

Teja (00:35):

Well, thank you for joining today. I’m excited to talk and learn about your role as a data scientist at Bridgestone, learn about how you fell in love with data science and all that stuff, but maybe a good place to start is just tell us about yourself.

Sabber (00:55):

Before even I tell like, about my profession at Bridgestone, I would love to tell like, my story: how I came here, and how I kinda transition my from career, like, from a different subject to a different kinda area. So I originally come from Bangladesh, and as I said to you before, like, I’ve been living in Tennessee like, forever, like, since I come to the USA <laugh>, so you could say, me, I’m not a US citizen. I am a Tennessee citizen.

Teja (01:24):

Yeah, yeah <laugh>.

Sabber (01:26):

So I came to USA back in 2013. So I went to school at the University of Memphis, and I completed my PhD in computational geophysics. (Teja: Wow.) So my goal was to become a professor or teacher, because I love teaching, and still I love teaching, but I ended up in [a] totally different industry setting, but it’s a different story. So during my PhD, my goal was to kinda understand how earthquake happens. Like, what is the physics or mathematics behind this kind of rupture process or failure process when the earthquake happens. So whenever I wanted to study like, this sort of kind of things, I got introduced to machine learning, computational geophysics, and parallel programming, and this is how I can apply all of those tools to understand the earthquake. So this is how I got familiar about the machine learning and artificial intelligence. So after my PhD, I got offer from the University of Southern California for the postdoctoral kind of things, then I refused this offer, because I thought like, why don’t I go to industry first and see the real life implication, because if I go to like, industry, then I’ll be able to see whatever I’m gonna make model, or whatever, I’m gonna make something, I would see the immediate impact to the customer’s experience or customer’s kinda life, right? (Teja: Mmm <affirmative>.)

Sabber (02:51):

I thought like, okay, let’s give it a try for a little bit, and if I still do not like it, then still I have the chance to go back to the postdoc or academia, something like this. But it turned out that okay. I like more corporate world and seeing the immediate impact, and then I stayed there in Asurion, and I worked there for like three and a half years as a data scientist. My job was to kind of develop a machine learning solution to detect like, real time fraud, because Asurion was the like, insurance company, and they [got] like, ton of claims every day. (Teja: Yeah.) One of the challenge was kind of like, they wanted to automate the entire fraud detection system from rule based to the machine learning based solution. So the machine learning scientist, my goal was to kind of develop a machine learning model that can detect fraud on the fly. (Teja: Mmm <affirmative>.) That means that’s the kind of real life challenge, right? (Teja: Mmm <affirmative>.) So yeah, so that’s kind of like, this is how I kind of transition from computational geophysics to the machine learning world. Then I never kinda looked back <laugh> to this academia and never went back to the <inaudible> community.

Teja (04:12):

That’s cool. So prior to starting the PhD program, did you have any familiarity with like, some of these tools or was it the first time?

Sabber (04:25):

So never ever, but I was familiar with some sort of like, programming, because even from my childhood, (Teja: Yeah.) for some reason, I kinda liked programming a lot. Like, when I was a kid, like, I used to learn like, visual basic. You know, visual basic used to be like, hot language, like, back then in 2005 or 2003, right? (Teja: Right.) Then I went ahead and kinda learned by myself, and I built like, small software, so this is what I learned, but I didn’t have any idea about machine learning or software engineering.

Teja (04:55):

How long did it take you to feel comfortable like, wrapping your mind around machine learning tools, just in the context of like, your PhD program? Couple years?

Sabber (05:05):

I think it was one of my professors who reached out to me and told me, “Hey, Sabber, would you be interested in working with me on a project that requires some sort of machine learning knowledge?” At that time I had very zero knowledge. Like, I didn’t have anything other than geophysics. Then I tried to learn, like, I think it was pretty easy for me to pick up all the kinda mathematics and physics, because in geophysics, I got introduced to all the kinda mathematics equations and everything, but whenever I got introduced to machine learning, I thought it’s gonna be hard, but didn’t take me that much time, because I had some prior knowledge.

Teja (05:45):

Yeah, yeah. On the math principles…

Sabber (05:47):

On the math principal and like, SVD or principal company analysis, I already use those techniques in geophysics, (Teja: Right.) but in different contexts. So it didn’t take that much time to pick up all of these kind of things.

Teja (06:04):

Do you find like, the context of solving business problems, let’s say, in fraud detection? Like, is it as intellectually engaging for you as, let’s say, studying or applying it in the geophysics context, like studying earthquakes?

Sabber (06:21):

Absolutely. I think, so there’s a huge difference between corporate problem solving versus academic problem solving. (Teja: Yep.) So in corporate means, industrial problem solving is kind of more of like, real life impacts. You have to think a lot of things, but in academia, you don’t have to, not necessarily, but you have the option to think about, for example, whenever you detect fraud, you have to think about customer experience. You have to think about companies saving money or company’s revenue. So for every false positive we make, that means we are giving hard time a good customer, right? So we falsely identify that, okay, a real customer is a fraudulent customer, which is kind of bad, but in academia, I didn’t have to think about this, not necessarily. (Teja: Yeah.) So that’s the kinda difference. So with this difference, it took me a long time to realize that academic problems versus industrial problems is kind of different. So even though theory behind this, everything is almost the same, so I would apply the exact same algorithm, exact same theory in different places, but context is different.

Teja (07:35):

Do you find that in like, a corporate problem solving context like, there’s the same level of statistical discipline, or is it different like, in terms of, let’s say, having the right sample size in terms of like, having those principles at play? Or are you guys optimizing for maybe something different like, hey, let’s get to some directional truth faster, and let’s make a decision quickly?

Sabber (07:57):

Yeah, absolutely. Like, the corporate or industrial problem solving is totally different, because as I told you before, you have to think a lot of things. So for example, say Asurion problem, Asurion is an insurance company, and that means it’s a financial company. That means we have to follow loss of regulation, (Teja: Yeah.) and if I even want to make a model based on neural network, I’m not able to do this, because neural network is a black box. So I’m not supposed to make any model based on neural networks. (Teja: Interesting.) Yeah, and not only this, like, they want to make a model that can detect fraud on the fly. So that means I’m limited [in] what type of algorithms do I have to make. So that means I’m not gonna use any algorithm that’s gonna take [a] long time to infer. So that means I have to use like, a lighter algorithm compared to the heavy algorithms. Even though those heavy algorithms…heavy means, like, that’s gonna give you long time to give you the prediction.

Teja (08:59):

Gotcha. So you have to trade maybe some degree of accuracy for speed, especially. Okay, gotcha. Okay. That’s really interesting. So what brought you to Bridgestone, and what’s the role like there?

Sabber (09:13):

<Laugh>. Yeah, so Bridgestone is kinda interesting and totally different domain. Like, insurance to the insurance, one spectrum and other spectrum is that really automated. As I told you before, underlying principles or underlying machine learning algorithms are almost the same, but you apply the same algorithm in a different domain. So the reason I came to the vision, because I wanted to see the different area, and that’s why I kinda thought, okay, I think it’s a good opportunity for me to jump, right?

Teja (09:47):

Nice. So in between Bridgestone and Asurion, you were at a startup, right? (Sabber: <Laugh>. Yes.) Tell us about that. How was that like?

Sabber (09:57):

So this startup actually is my startup <laugh>.

Teja (10:00):

Okay, awesome. Yeah.

Sabber (10:02):

Yeah. So I started working during, actually back in 2015, when I was a creative student. So being an immigrant and being a non-native English speaker, I found that, okay, so there is no place to find common ground people. Like, for example, I wanted to find somebody who can talk in my way or, or see the problem in my way, because I’m totally different, from different culture, and whenever I come to the university, like I see I’m an outlier <laugh>, even though they’re a ton of like, people from my society or my culture, but still, I feel like I’m one of the teeny tiny person from different society. (Teja: Yup.) It’s kinda hard to match. So this is how I kind of came to the idea of Xoolooloo, that I established. So the goal or the idea of Xoolooloo is to find people based on your very niche interest and uncommon hobbies, <laugh>. So for example, I’m a chess player, and I’m also interested in algorithmic trading. How do you find people who have these two kind of interests? Like, traditional search engines like Google or Bing, they wouldn’t help you to find those people nearby you, because the search engines are not meant to find people (Teja: Yeah.) based on your geography. (Teja: Yeah.) Neither the traditional social media like Facebook or LinkedIn. I do not put like, I’m a mushroom grower <laugh> or I’m a Ethiopian food lover. I do not put this interest in my LinkedIn profile or Facebook profile, right?

Teja (11:49):

Yep, yep, and even the like, dating apps or whatever, (Sabber: Dating apps, yeah.) They’re not set up to have these dimensions exposed, then (Sabber: Exactly.) construct the search off of these. Yeah, totally.

Sabber (12:01):

The interesting thing is like, if you see a data scientist, for example, in a conference, and at the same time, you see this, okay, this data scientist happens to be an algorithmic trader, and this data scientist happens to be a chess player, you get more excited. (Teja: Yeah.) That means you get more excited on your hobbies rather than your primary skillset, right? (Teja: Yes.) So that means there is some psychology [that] happens all the time. Like, whenever I see people like, with common interests, not my primary skillset or primary interest, but my secondary shared interest, that kinda tripped me. Okay, so why don’t you kinda make a platform and work on it? So this is how I kinda came up with the idea, and this is how I started working in my PhD and continued through <inaudible>.

Teja (12:55):

That’s awesome. So how’s that going? Are you having fun building it?

Sabber (13:00):

<Laugh>, Honestly, yes and no <laugh>. So yes means, because the problem, I think it’s not my problem. I see whenever I talk to people, it’s the problem across like, different demographics. Whenever I talk to the older people, it’s very cool, because older peoples are the one, they kinda live by themself alone and they want to find some other people based on their niche interests so that they can talk to each other, they can go meet in person, but whenever I talk to the younger people, they also see the exact same problem to the different point of view, but exact same people. Like, whenever they come to this school, they want to find some other students with the same interest. Not [that] they’re lonely, but they want to find the same common ground people. (Teja: Yeah, yeah, yeah.) So, it’s fun, but it’s not fun that I’m working by myself, and I’m thinking that probably I’m just gonna quit on this idea and start something else <laugh>.

Teja (14:02):

Yeah. It’s interesting. It’s hard to build a product and a company, you know? It’s not easy. What do you like about kind of building your own product, and how is it different maybe from day-to-day life, like, in the corporate environment?

Sabber (14:17):

That’s a pretty good, good question. [It’s] interesting, because whenever you make your own product, you have the 360 degree freedom to make your own decision [about] like, how you want to pivot or how you want to go farther along this road. Not only this, like, you have the total control, and you can make decisions very fast, very quick, versus you have to depend on your leaders’ <laugh> decision in the corporate world, and you have to work on the project that you are assigned to in the corporate world. Another kind of good thing I enjoy, especially in startup kind of thing, like, you get to see the people’s feedback, and you can fine tune your product, getting your feedback, getting feedback from the customers, but you don’t have that luxury in the corporate world, because the project is fixed and project has some fixed objective and goal already defined, right? Yeah, there are some startup like, projects in the corporate world, but still it has to depend on your leadership voice or something. Right? So I am a startup person, and I like [to] build products no matter what. Like, whenever I get time, I just do this for myself <laugh>.

Teja (15:42):

That’s true. The freedom is awesome, (Sabber: Yes.) you know, but you know, it’s also like, you have to be disciplined to say “no” to things too, right?

Sabber (15:51):

Absolutely. Yes, yeah. So there’s a trade off all the time. Like, there’s a trade up in [the] corporate world, there’s a trade off in the startup world.

Teja (15:58):

Yeah, totally. Do you think like, let’s say, what do you find more like, intellectually engaging? Do you find like, solving like, big problems at a big company where you’re well resourced and you have a team to support you, do you find that more stimulating, or do you find like, being able to be really hands-on, develop the product, talking to customers? Like, do you find that stimulating like, at a smaller org or product?

Sabber (16:25):

The latter.

Teja (16:26):

The latter? Okay, yeah. Fair enough. Why?

Sabber (16:29):

Yeah. Yeah, by heart I believe [that] I’m an entrepreneur since my childhood. (Teja: Yeah.) Like, even before I come to USA, I had a small organization in Bangladesh that I run myself. So I established this. It’s kind of like, some financial kind of thing, but I founded it myself, and I had like, couple of users, but I had to leave and come back here, and then when I go to school, I started Xoolooloo. Now, even I joined to Bridgestone, I always ask my boss or leader that, hey, if you have a new project, it’s very new and you want to give it a try, see if this is going to be successful or not, then just assign to me, not the project you already have <laugh>.

Teja (17:16):

Yeah, no, totally. I mean, I feel like, you know, if you were to take like, the average person, people who immigrate to a new country are definitely like, in the top quartile of like, risk appetite, right, just generally speaking. So I feel like, you know, so it’s funny, I talk to my parents, and they’re like, “Oh, it seems like, really risky to like, start a business.” I’m like, “Well, you guys moved to a new country without really speaking the language.” Like, that’s a substantial amount of risk, maybe even more risk than I would take, you know? (Sabber: Exactly, right.) So, yeah, it makes sense to me that you feel like the entrepreneurial path is like, exciting and that matches like, your risk appetite, for sure. (Sabber: Yes.) So you mentioned that you like algorithmic trading? Talk to us about that. Like, what do you like about it, and you know, what are like, what are you working on there?

Sabber (18:12):

<Laugh>. So it’s funny. Like, I’m the people who is interested in everything <laugh>. (Teja: Yep. Same.) So algorithmic trading is one of the niche area that I found. Okay, probably down the road, I will be working on this area. So, okay. So I’m a very stock market fan, and I love to study, and I have investment, some stocks, and I try to follow stock market, or market trends, or these kinds of things, and I trade and I used to be a day trader <laugh>. (Teja: Nice.) Of course I lost a ton of money, and I earned [a] ton of money from the day trading, and I feel like, why don’t I kinda automate everything, because now that I got my hands on problem in like, manually trading and everything and everything, so I feel like there is a pattern in trading. (Teja: Yep.)

Sabber (19:08):

Not pattern in kind of like, when the market is going to up and down, but there is a pattern in trading, (Teja: Right.) and machine learning is all about kind of pattern recognition, right? So I thought like, why don’t can I use my skillset into a trading problem and see if it is helpful or not? So this is where like, [I] started working on algorithmic trading, and building my own algorithms, and testing it [in] real life. I haven’t tried my algorithm in real life yet, but still like, I’m just kinda making [an] algorithm and testing it myself at home, and I have a deep learning GPU workstation at my home. So this is a good advantage. This is a very good thing. So I use it all the time.

Teja (19:53):

How does like, it work to set up like, a robot to trade for you? Like, do these trading platforms have like, APIs that you can access? 

Sabber (20:02):

Yes. Oh, yeah, of course.

Teja (20:04):

Interesting. I had no idea, ’cause people would talk about it, and I never really like, oppressed, you know, ’cause my strategy is I just put my money in Vanguard funds, and I leave it alone. You know, that’s…<laugh>.

Sabber (20:20):

So algorithmic trading is kind of like, more of like, dynamic, like, whenever [a] market changes, (Teja: Yeah.) your algorithm [will] act on it. So for example, so say Apple, okay? (Teja: Yeah.) So your favorite stock is Apple, and you wanted to trade on Apple, then how do you do it? So for example, today, Apple announced that they’re gonna have some groundbreaking product, right, and of course the market is gonna react very positive, right, and I want to buy right at this time so that my Apple stock gets better. So how do they do this? So you automate the entire kind of thing. So you have your algorithm that listens to your market, (Teja: Right.) and you have an algorithm that listens to your news kind of thing or portal everything, and your algorithm also knows like, past pattern of how Apple can have been working in past like, two months or last one month. So you combine all of this kinda information, and then your algorithm makes a decision. Is it the right time to buy, or is the right time to sell? (Teja: Right.) So this is all about an algorithm trading <laugh>.

Teja (21:30):

That’s cool. That’s cool. Yeah, I had no idea that you can just access like, these trading platforms’ APIs and make…

Sabber (21:37):

Oh yeah. Yeah, yeah. There are a ton of platforms out there like Alpaca, Webull…they just kind of give you access to the data.

Teja (21:46):

Yeah. That’s cool. Do you like reading biographies of people?

Sabber (21:52):

I used to, but I read a book…so currently I’m just reading a book. I’m not quite sure if you’re seeing this. It’s called the Prisoners of Geography. Tim Marshall is a very good writer.

Teja (22:02):

Cool. Okay. I’ll have to check that out.

Sabber (22:04):

Yeah. So this book is about like, geography, and a country, and the power <laugh> in the entire war in politics.

Teja (22:16):

Yeah. That’s interesting. I read a book a long time ago called Guns, Germs and Steel. (Sabber: Ah.) If you like books on history and different ways to think about like, the modern world based on like, I don’t know, path dependence from history, things like that, that might be a good one too, but I will check out this book.

Sabber (22:39):

<Laugh>. I love this book, yes.

Teja (22:41):

Yeah, that’s sweet. The reason why I ask is I wonder if there are like, some traders that you follow that you like, you know, anything like that.

Sabber (22:52):

So yeah, so there is a one guy. He lives in…so he’s our generation guy, and he’s way, way more better than me. So I usually, like, his name is Ricky, so he’s based in, I believe Arizona somewhere in the desert. So he’s not a big like, guy, not like Warren Buffet, but he’s a day trader and successful day trader.

Teja (23:20):

Yeah. How much capital do you need to have allocated to day trading to generate sufficient returns to justify the mental time investment in that?

Sabber (23:31):

Well, it all depends <laugh>. So for the early day trader, like, for example, if I’m very novice and new here, I wouldn’t recommend to invest much, but it all depends on your experience in something. Even [an] experienced day trader will lose a lot of money. (Teja: Yeah.) Millions and millions. (Teja: Yeah.) So it’s all about kinda experimenting on trial and error. Like, see what does work and what does not work. (Teja: Yeah.) It’s all about emotion.

Teja (24:03):

Yeah. Yeah, that’s true, and managing that emotion, right? Yeah.

Sabber (24:07):

You have to manage your emotion. (Teja: Yeah.) If you can do this, I think you could be a very good millionaire based on your trade.

Teja (24:14):

Yeah, yeah. That’s cool. Also I think it depends on like, your day-to-day job opportunity cost, right? Like, if you have a lot of opportunity costs, then you have to risk a lot of capital and day trading to justify like, the marginal investment in day trading versus your job. That’s something that I think about too. (Sabber: Absolutely.) Yeah. Yeah, well, you have a big nut to cover, yeah. That’s cool. What would be your advice for, let’s say, fellow data scientists that are looking to, you know, move up in their career and let’s say have like, this panoply of interests, and like, because I feel like most data scientists are like, math nerds, and if you’re a math nerd, you probably are into trading, right, and you’re into probably building like, your own thing, ’cause like, you’re probably sophisticated and interested in like, leveraging your knowledge, right, in those ways. What would be your advice to other data scientists in managing their career versus other passions and professional interests that they may have?

Sabber (25:19):

I think that’s a pretty good question. Like, my one particular advice to the fellow data scientists is to kinda…if you really want to pursue in data science or machine learning, I’d highly recommend you to know the underlying roots of any algorithm that you work. So for example, if I’ve been interviewing a lot of folks like, in my past careers, but I see they lack like, underlying foundations, which is kinda…because if you have the good foundation, you can make whatever you want. (Teja: Right.) Potentially it’s totally endless. So even if you do not have [a] job, at least you can make a lot of money without even [a] job, (Teja: Yeah.) just because you have the foundation, and you can treat different problems in different ways, right? So my advice is to kinda know better, like, whatever you are trying to working on.

Teja (26:16):

Do you mean the foundations as in the math, like, the math principles or the tooling?

Sabber (26:22):

Both. Like, math, because the tool is also important too. Like, you have to know the tools, because you’re gonna use it. (Teja: Right.) Different tools has different underlying principles. (Teja: Right.) So you not know the principles, then you do not exactly, basically, know nothing about anything.

Teja (26:38):

Yeah. Yeah. Totally. Awesome. Where can people find you, or your startup, or whatever you’re working on next on the internet?

Sabber (26:48):

So, yeah. So my next journey, or my main next kind of thing would be probably build startup or build anything based on LLM or machine learning, because I believe this is where my expertises are in, so then why shouldn’t <inaudible>? (Teja: Right. Yeah, totally.) I like the LLM concept and <inaudible>.

Teja (27:11):

Yeah. So people can search you on LinkedIn? Nice. Do you have a personal website or portfolio that you want people to go to?

Sabber (27:21):

Unfortunately, I don’t have any personal website, but LinkedIn is the only one that you can reach out to me.

Teja (27:26):

Awesome. I think the community will really love this one, because I feel like, you know, (THE FRONTIER THEME FADES IN) a lot of people at least tell us like, “Hey, we want to hear from people who like, are working and building something on the side, and thinking through, and maybe just, you know, have a little bit of experience in pushing their career forward.” So yeah, this will be a good one. So appreciate this, Sabber. Thank you so much for your time.

Sabber (27:52):

Absolutely. Thank you.

Abbey, via previous recording (27:53):

You’re listening to the Founder to Founder podcast, powered by Gun.io’s Frontier Network. We release a new episode every Thursday morning, so be sure to subscribe on Spotify, Apple Podcasts, Stitcher, or wherever you stream your music. Please leave us a review and share with your friends. You can follow us online at the Frontier Pod or drop us a line at [email protected] to get in touch about hiring world class tech talent.

(THE FRONTIER THEME ENDS)