Overcoming Cultural Hurdles in Tech

As the first Mexican woman to get a PhD from Stanford, Debbie Berebichez has experienced what it takes to challenge norms and boundaries. She speaks about her personal experience overcoming cultural norms regarding women in STEM and her professional experience training company culture in data science.  

Debbie Berebichez: It’s very important to encourage people to be evidence based. To see, okay, if you have a new idea for business, search for the metrics that are going to tell you that that idea may work or why it may not work, but set those parameters beforehand.

Ginette: I’m Ginette,

Curtis: and I’m Curtis,

Ginette: and you are listening to Data Crunch,

Curtis: a podcast about how applied data science, machine learning, and artificial intelligence are changing the world.

Ginette: Data Crunch is produced by the Data Crunch Corporation, an analytics, training, and consulting company.

Up until this month, Debbie Berebichez was Metis’s chief data scientist. We talk with her about her journey into STEM and her perspectives on data science.

Debbie: Thank you for inviting me, Curtis. I love your Data Crunch Podcast, and it’s an honor for me to be here.

I have kind of a unique story in that I was born in Mexico City, and I grew up in a fairly conservative community that discouraged girls from pursuing careers in STEM and specifically in science and math. And I was a very curious child. I, I asked a lot of questions, and I wanted to know things about the world and, and understand about the planets and how things work and growing up in a community like that, you know, it was cherished, but at a certain point when it came time to go to college, my teachers in school, as well as the counselors and my friends, including my parents, they were like, “no, you probably should pick a more feminine career, something easier.”

My mother told me, be careful. Don’t tell boys that you like math ’cause you may end up not being able to get married, which almost happened. But, uh, it really, it was just a thing that I had to hide. And with that came a lot of insecurity because I just thought I’m not good enough in math and these topics, and, and I will never, never be able to do it. Until the day came to go to college, and my advisors in school had said, “you know, what, why don’t you study philosophy?”

The funny thing is during high school, when people were doing their crazy rascal things and whatnot, I was actually getting books from the library about obscure physicists, like Tycho Brahe, and reading about how they were locked up in a tower or an observatory, and I thought to myself, “you know, maybe I’ll be like them. Maybe I’ll be locked up in, in some tower and alone, and, and, you know, not be a very sociable person. However, I’ll have my science and my observations with me, and that made me happy.” And so I grew up like that.

And so when I had to decide to go to college, someone said, “you know, philosophy also studies several people and what their, their ideas about life are, and they’re very curious.” So I said, “okay, fine.” That, that appeased everyone around me. And I started studying philosophy. Two years in, my hunger to know about the world and the universe was louder than ever, and it was not going to go away.

So I decided to apply behind everyone’s back to schools in the US, because I had learned that in the US, you can do a double major and study more than one topic, which I couldn’t do in Mexico, but I was afraid because my parents couldn’t pay an American University when we were paying an eighth of that in Mexico City for a private university.

So I didn’t know if I could afford it in the middle of the application process. I got a beautiful offer from Brandeis University, a small university in Massachusetts that offered me a full scholarship that was offered to two international students per year to attend Brandeis. And I was so incredibly lucky and happy.

I flew to Massachusetts. I had never seen the snow before. It was in the middle of the winter, ’cause I was a transfer student from Mexico. And here I am, I enrolled in my philosophy courses, and they didn’t really know what to do with me because I had already studied so much in Mexico, given that that was the only topic we were allowed to study.

And so, my first semester, I had the courage to take a very generic course in astronomy, Astronomy 101. And I met the, a graduate student who was the assistant for the class. His name is Rupesh Ojah, and he came from India and Rupesh and I became very good friends, and we would walk around campus and I would ask him all kinds of questions about the universe and planetary emotion and the laws of physics.

And he was the first person to really believe in me. And he said, “you know what you, you’re not the typical student that just has, you know, the, the thirst to have an A in, in the homeworks. You really care about this. You have so much passion. So one day we were walking in Harvard Square, and I told him, “Rupesh, I just don’t want to die without trying. I don’t want to die without trying to do physics.”

So he got up, and we called his advisor, who was the head of the physics department at Brandeis. We had a meeting, and he basically handed me a book. It was calculus in three dimensions. It was called Div, Grad, and Curl, which was an alien language to me. And he said, “look, there’s somebody else.” ‘Cause I had a problem that my scholarship was only for two years, and that’s what I had left.

Just those two years. And so to put a whole physics major when I was not confident and I had very little math background was going to be a big challenge. And so Rupesh, and his advisor said, “there’s someone else who has done this in the past. Edward Witten, he’s the father of string theory.” I thought they were pulling my leg, but comparing me to him.

And they said, “well, uh, we’ll let you skip through the first two years of the physics major, if you’re able to cram all of these topics for an exam in two months at the end of the summer.” So Rupesh decided to devote his entire summer to tutoring me and mentoring me. And it was incredible.

And the reason why I’m sharing this story, Curtis, is because I always wanted to pay Rupesh for all that he did for me. And he said to me that when he was growing up in Darjeeling, in like the, the mountain in India, there was this old man who used to climb up and teach him and his sisters, math, English, and the tablet, a musical instrument. And when Rupesh his family wanted to pay him, the old man insisted and said, “no, the only way you could ever pay me back is if you do this with someone else in the world, and that’s how my mission in life began to encourage and inspire other people, especially minorities or women who like myself feel attracted to STEM, but who, for some reason, feel that they cannot achieve their dreams.

So after finishing Brandeis, I was accepted by the then Current Nobel prize winner, Steve Chu in physics at Stanford. And it was just incredible. This person that two years ago knew very little. I mean, even algebra was rusty was all of a sudden accepted to Stanford. So I became six years later, the first Mexican woman to get a PhD in physics from Stanford. And that’s when I realized that I had a responsibility to spread my message and help others. And since then, I, I, I did two postdocs in applied math. And physics at Columbia University and NYU at the Grant Institute. And I’ve just been working with numbers all my life. And I’ve had also these science communicator career on the side.

So for example, right now, I’m the chief data scientist at Metis, which is a data science training company. And we do bootcamps, and we train corporations and businesses in, in data literacy. And in, in, you know, how to increase business insights through, through data literacy. And at the same time on the side, I, I, I’ve been hosting, co-hosting a TV show for the Discovery Channel called “Outrageous Acts of Science” that helps me exercise. The explaining of complex concepts in lay and entertaining ways.

Curtis: And that’s, that’s an inspiring story. And how did you, I’m also interested in the transition you had between physics and into data science, right? And so a lot of, I mean, getting to physics is amazing. I think a lot of people would share your sentiments about, you know, some of these challenges that you had, but then there’s also this point where you’re now doing data science and data literacy, which no doubt, your physics PhD helps in that. But it is, uh, correct me if I’m wrong, but it’s, it’s probably a little bit different than if you had, you know, stayed doing research in physics. And so, so how did that transition work?

Debbie: Absolutely. I completely agree with you. It was not easy. It was definitely a challenge. So the first thing is that physicists who couldn’t find jobs, which were many, many, especially after the Cold War, a lot of the departments were shrinking in the US and whatnot. So a lot of them found their way to Wall Street. So when I was looking for a job 15 years ago, they, Wall Street had a very kind of open policy where they would go to physics graduate students and interview everyone.

And we were just put on the hat and the title of quant. And that’s what I did. I applied for a position in academia. I wanted to be a professor at some point, but then I realized after my postdocs that, that it could be very isolating, and it wouldn’t make me very happy.

So I decided to give it a chance, and I became a quant. And so I worked in risk analysis for both a hedge fund named AQR, and then for Morgan Stanley Capital International, and I was building risk models and, and explaining them and selling them to hedge funds and banks. And yeah, I realized that what I was doing was data science, but it was only a very narrow area of data science, which was time series and working with numbers, not with images and not with audio or video, but it was certainly, you know, we were building models and we’re using data. And I realized that Wall Street, uh, I spent six years in Wall Street and I was into some of the intellectual challenges of building the models, but just being interested in money and how the stock market is doing was not my thing.

Like I, I, as a child, I was very curious about the world. I was still very curious then. And so I decided to switch, and I had heard by this time I was friends with Hillary Mason and others. So at that time I realized. There’s a huge field that’s kind of being born, or it’s, it’s starting and it’s, there’s a boom for data scientists, so let me give it a try. And I remember I attended Strata, the conference, at one point, and they recorded me on video, and I, I’m embarrassed to say that I, I said something like, “well, there’s nothing new in data science. Physicists and Wall Street people have been doing it for 50 years and this and that.

And then boy was I surprised when I actually ended up taking, taking a data science course. It was a four month evening course with General Assembly to basically be able to translate my mathematical and computational skills into data science. And it was really difficult. I mean, there were things in physics, for example, we do statistics, but nowhere near the depth that you need to know in order to do data science, and, you know, there were linear algebra and calculus and all those of course I knew by heart and they were advantages, but there were many, many other things like working with them.

Images or, or, or words with NLP and all that, that conceptually were very difficult for me to understand. And, and what was accuracy about in a model and, and whatnot. And so after doing that, I got my first job in traditional data science. Uh, ThoughtWorks, which is a boutique consulting company that mostly does software, but they were starting their data science team.

And from there, I realized that I wanted to combine data science with teaching because I really missed that academic field where I could be part of creating a curriculum and, and really kind of effecting change in, in people’s brains about how the world works and how we are inundated with data and what insights you can gain from it.

And so that’s when, at another strata, my friend, Cathy O’Neil. So Cathy and I were speaking, and I said, I’m looking for another gig, but it has to be something really special where I can really have an impact. She introduced me to Jason Moss, who is the president and CEO of Metis, which is where I work now.

And Jason hired me to sort of be his right hand. And since I joined Metis five years ago, I’ve created curriculum for both universities, Dublin business school, for example, as well as for the bootcamps and for corporations. I’ve been doing thought leadership. I’ve managed the team of science, uh, data, scientists, and instructor. So it’s been an amazing ride.

Curtis: That’s awesome. Now I’d like to, I’d like to dive into sort of the subject matter there, the data literacy. But before we do that, I’m just curious, because I imagine some other people may have this question that is how, how does it feel now doing more data science type work instead of physics being, being that maybe physics you could say was kind of the first thing that really interested you like about the world and this kind of stuff.

And now you’re doing more data science. Is it as enjoyable to you? Is it more enjoyable or maybe there’s not a comparison there, but just people that maybe are making that transition.

Debbie: Absolutely. Well, I, I’m going to be honest with you. I think I will always miss doing basic research because if you have that curiosity from a very young age, I don’t think it ever goes away. My husband is actually a physics professor, so we have endless discussions about physics and what are the new discoveries and whatnot. So I keep my mind in physics, somewhat. But on the other hand, I do see that people who become professors and stay in academia end up spending quite a bit of time applying for grants and doing a little bit of department politics.

So, you know, it’s not as idealized or as pure as when we are undergrads and we just spend all the time in the lab figuring things out. And so from that perspective, I do think that my personality lends itself better for working in business. And so, you know, in data science is such a vast field that you can always find ways of contributing to different projects. You can increase the literacy of a hedge fund, for example, and then seeing the aha moment, and they’re like, “okay, I no longer have to use Excel for this. And I have a Python code that allows me to automate this instead of taking four hours, it takes four minutes. That’s really cool.”

But also being able to help, like at Metis, we did something called Metis for Good, where we, I took all the alumni from Metis, and we have tons of amazing, amazing alumni from our bootcamps, and I wrote to them right after an earthquake that happened in Mexico about four years ago. And I said, you know, “we, I need to build a map in real time showing the data off of where, what things are needed, where.” And a bunch of alumni helped me, and we built it on, it was shared. I forgot how many hundreds of thousands of times during the earthquake and the fact that we were able to help was just incredible. I think, you know, MediSYS corporate training has had the opportunity to train businesses in sort of taking their insights to the next level and getting data literacy to spread across the company, which is my sort of my big thing that data should not stay in silos and only the technical people comprehending what’s going on, but it should be adopted by every single person, HR, you know, the, the chief executive officers. Everyone should have a stake in the insights that are being gained by data analysis.

So that has allowed me to not discover things about the universe, but definitely discover things about the world and about how people behave and how, you know, different projects evolve. And that has been equally fascinating.

Curtis: And let’s talk about that a little bit ’cause a lot of companies, as you know, are struggling with this, how do we increase the data literacy of our employees?

And maybe more than that, how do we put data literacy to good use so that we can make better decisions and these kinds of things. How do you approach it? How do you think about it? What are some of the key points do you think?

Debbie: Yeah, so I think what’s happening is that data science as a business term is maybe 10 years old or even 15 years old. And so a lot of companies have already, uh, the novelty has worn off, and everybody’s like, “okay, so what can we do with it? It sounded so promising, like the Holy Grail, like it was going to save my PNL and everything was going to get better after I hired all these data scientists, but nobody ever communicated to those stakeholders what data science actually is, what its limitations are, what type of people should be working on data science. What kinds of data are useful? What kinds of problems can they solve? And what. The company actually enacts some of the decisions that the insights would bring about.

And, and so people in industry have been frustrated, I think, with these new sexiest job of the 21st century, like Harvard business review called it. And so I think because of that, we need to sort of go back a little bit and redefine what data science is by educating every single person that works in a company. And by this, I don’t mean that everyone’s going to have to become a programmer and a data scientist, not at all, but everyone should at least have access to the ideas, to the goals and to the analysis that’s driving the decisions that the company makes.

So say for example, that HR wants to know why women after maternity leave, tend to not come back to their executive positions in that company, what can they do to support them? And they try different measures and programs and they have data about it. Well, if I only have a bunch of, you know, technology people in the IT department analyzing that data, we’re probably not going to get very deep insights because we have to bring all that data to the HR department and to the women in the company who are stakeholders, because this is about them. And so when you, you allow everyone to gain access to the data, the graphs, the charts, and you collect information and feedback, then everybody has a stake in what you’re trying to accomplish. And so you’re much more aligned with the success of the program or whatever you’re you’re implementing. And when we see companies that have done this, we see, you know, a hundred percent difference in our corporate training.

On the other hand, we’ve seen companies where the chief executive office doesn’t even agree on why data science is important or data literacy, and they hire a, a vendor or somebody to train their, their technical team.

And then at the end of it, the technical team is trained, but they have no idea why. And why they should use it. And so every, even though they made a big investment, the, the data more data literacy did not get them anywhere. And so that’s why it’s really important to explain it, to talk data, to brief data and to have it as part of the, sort of like the blood system.

That’s like spreading blood everywhere in the, in the company the same way we want to spread the message of data.

Curtis: It’s a tall order. And, um, I think there’s a lot. I mean, what you’re talking about it, if you, and obviously we can’t boil this down to one or two things, right. There’s so much here, but, but maybe if you had one piece of advice or two pieces of advice that you think are maybe the most important that someone could listening to this episode today, they could take it and, and put it into action and it would help them. What, what would you say?

Debbie: I would say my first piece of advice is critical thinking. Uh, Richard Feynman, a very famous physicist, used to say that it’s very easy for us to get fooled by, by what we see out there. But it’s actually a lot easier to fool ourselves. So when we are working in business or when we’re analyzing what our newspaper said about politics or COVID or whatnot, we tend to bring our own biases.

And it’s very important to encourage people. To be evidence-based. To see okay, if you have a new idea for business, search for the factors or the measurements and metrics that are going to tell you that that idea may work or why it may not work, but set those parameters beforehand. Don’t say, “oh, my business idea was successful,” because three more people adopted your app or because you have to set the parameters of how many people should adopt the app for it to be successful before.

And so to educate people on critical thinking and evidence-based approaches I think is, is like where I would start. And of course they’re practical, very practical examples of how to do that and how to teach people how to read charts.

I actually teach a workshop on statistics and the art of deception, which is a very down to earth course, very little math involved, but I go through a lot of graphs that try to mislead the public in how, you know, the conclusions they make them sound seem bigger than they actually are. So if you teach everybody in a company to read graphs and to question the data who has the data, why, what for et cetera, I think your employers will be, employees will be much better prepared for dealing with company issues and projects.

Ginette: A huge thank you to Debbie Berebichez. Feel free to reach out to her on LinkedIn or Twitter. As always check out our transcript and attributions at datacrunchcorp.com/podcast.



“Loopster” Kevin MacLeod (incompetech.com)

Licensed under Creative Commons: By Attribution 3.0 License