Women in Data Science Worldwide

Lesly Zerna earned her undergraduate degree in Telecommunications Engineering at the Bolivian Catholic University and then traveled to Brussels to complete a Masters in Computer Science. After returning to Latin America, she began teaching data science and AI both in universities and virtual platforms and today her courses have thousands of online students. She brings insights from her experiences working in large companies overseas to her students in Latin America. 

For those just starting in data science, she says you must first identify your personal learning style (e.g., visual or text) to improve your learning experience and start with a general overview of the field. Next, find a practical topic you’re interested in, and look for projects, examples, authors, researchers who are working in that area. Do all of this while continuing to develop the fundamental skills you need (e.g., languages, platforms, frameworks) in data science. 

Lesly transmits her passion for learning to her students by using real scenarios instead of theory in textbooks. She lets them experience what works, shows the development process, and where common mistakes are made. She says it’s important for students to find where the problem is, know how to solve it, and make decisions. She believes there’s a lot to learn from the world of entrepreneurshipas you not only develop a project, you also have to develop the skills to explain and present the project, sell it, and negotiate. 

She believes that mentoring is essential to break down barriers for women. It can help dispel myths and biases about women in science and technology jobs, and learn from successful women that in spite of a hard path, they were able to achieve and follow their dreams.

RELATED LINKS
Connect with Lesly on LinkedIN
Find out more about the Universidad Privada Boliviana
Connect with Cindy Orozco Bohorquez on LinkedIN

Leda Braga is the founder and CEO of Systematica Investments, a hedge fund that uses data science-driven models to support its investment strategies. Leda was born and raised in Brazil and found her way into the financial sector after getting her PhD in engineering and spending several years as an academic. 

Her financial career started with seven years in investment banking at JP Morgan and then she joined the hedge fund startup BlueCrest in 2000. She explains that while her funds did very well during the 2008 financial crisis, the time felt like an existential crisis because you didn’t know if the major investment banks were going to survive. But she said it was a formative time and she learned many lessons. Several years after the financial crisis, she spun off her own firm, Systematica Investments focused on systematic trading.

Leda explains that systematic investment management is data science applied to investment. The systematic approach makes the investment process less reliant on the random nature of forecasting and more reliant on risk control in portfolio construction.

Both discretionary traders and systematic traders are looking at information to try to make decisions. Those who do it on a discretionary basis tends to look at the data and make a decision to make money on a trade. Those that look at data on a systematic basis build data-driven processes for trading strategies for certain risk profiles and preferences that will produce consistent returns over time. She says the responsibility weighs heavily on her to ensure a good return because people's pensions are part of the money her firm manages.

While she believes strongly in the power of leveraging data science in investment, we’re not yet at a point where AI allows us to do “autonomous investing” because there's a large element of randomness in markets and relatively sparse data so learning algorithms have limited use. She says that the only way it might be possible is if you’ve compartmentalized and narrowed the scope to the extent that you have a controlled amount of randomness. Learn more about Leda and systematic investing in her 2018 WIDS presentation, When Data Science is the Business.

RELATED LINKS
Connect with Leda on LinkedIn or Twitter
Find out more about Systematica Investments
Connect with Margot Gerritsen on Twitter (@margootjeg) and LinkedIn
Find out more about Margot on her Stanford Profile
​

A Colombian engineer, Jessica is fascinated by the processes and complexity of water supply systems in urban areas.In her post doc research in Australia, she brings together her expertise on the water hammer and transient flow waves to create an AI model that is able to identify where pipeline defects are faster and more accurately than existing techniques.

She explains that in data science, the most important stage is understanding the problem. You need to bring in basic knowledge of the problem and expertise from other disciplines that are involved in a problem and combine that with artificial intelligence. AI is an important tool but just part of the solution. It’s critical to maintain all the legacy of knowledge and understanding of a problem. AI can make it simpler to apply, but you can’t leave behind the physics or knowledge of the hydraulic part of water movement. 

Working in industry, she has found that it’s important to first understand how the system works. In these large companies in charge of delivering water, each person has different objectives, so you need to understand how the company works, who is in charge, what are their objectives, and how they measure their success. If your research project aims at those things, they will be more receptive and a better chance of success.

Jessica has learned in both research and industry consulting that nothing works the first time and it’s important to not to let those little defeats build up in your head. You need to trust yourself. There are many moments in life when you are criticizing yourself, and you realize that the biggest enemy you have is yourself. She just breaks down the problem into small parts and then solves each part one by one. She is passionate about teaching and inspiring young engineers about the importance of water and the future of this invaluable resource.

RELATED LINKS
Connect with Jessica on LinkedIN
Find out more about the University of Adelaide
Connect with Cindy Orozco Bohorquez on LinkedIN

As a quantitative social psychologist, Karolina has always been interested in using data to measure human behavior to try to understand it better. She has researched questions around political attitudes and polarization, particularly in light of Brexit and Trump’s election in 2016. She wanted to understand how people could arrive at completely different understandings of the world and reflect it in their voting decisions. One of her findings was that in the American two-party political system, people tend to identify as either Republican or Democrat and are more likely to agree with statements from their identified party. People use identity cues as mental shortcuts to judge information because there’s simply too much information to decipher. She says the polarization is stronger in the US where there are just two major parties compared to other countries with more choice of multiple political parties.

After her undergraduate and Ph.D. in psychology and two post-doctoral positions, Karolina decided to leave academia and to work for the nonprofit Teach First. She felt there was a lot of pressure in academia to become an expert in one niche and she wanted the freedom to pursue multiple topics that interest her. 

When she landed her first job outside of academia, she said the adjustment was a bit challenging, for example, when she first got the data to work with. In academia she knew exactly what the labels were, but in a new organization, she had to figure out how they measure things, what information they store, or what they use as a proxy for a certain behavior. 

As a researcher at Teach First, a non-profit in the UK that trains early career teachers to work with schools in disadvantaged areas, she is currently evaluating the impact of their programs in schools across the UK. She wants to know if their programs actually have an effect on the pupils that are being taught by their teachers compared to others.

When reflecting on her career, she says there have been times when she questioned whether she had the right skills. She has learned that it’s OK to be uncomfortable in a new position. With any new challenge you take, it takes time to get to know that new environment, and get to a place where you can start confidently contributing. It’s part of growing and learning, the satisfaction that you get from crossing that bridge from being very unsure to getting to place where you're comfortable and succeeding is very rewarding. The process of maturing in your career is accepting that this is just going to be part of the journey.

RELATED LINKS
Connect with Karolina on LinkedIn or Twitter
Find out more about Teach First
Connect with Margot Gerritsen on Twitter (@margootjeg) and LinkedIn
Find out more about Margot on her Stanford Profile

EPISODE NOTES
WiDS Executive Director Margot Gerritsen welcomes her new co-host, Cindy Orozco, in a wide-ranging conversation about their career paths and valuable learnings along the way. 

Cindy is thrilled to be joining as podcast co-host and believes that showcasing women at all stages of their careers shows that we “share the same fears or experiences every day. It's just that some of us have been on the path a little bit longer than others.” 

Cindy is an applied mathematician who is currently working as a machine learning solutions engineer at Cerebras Systems. Originally from Colombia, she loved applied math, and did a master's in civil engineering and mathematics from King Abdullah University of Science and Technology (KAUST), in Saudi Arabia, and a PhD in Computational and Mathematical Engineering from ICME at Stanford. She met Margot at Stanford and has been contributing to WiDS for many years at conferences, workshops and datathons.

After answering some questions about herself, Cindy stepped right into her co-host role to interview Margot. A native of the Netherlands, Margot said her career path was similar to Cindy’s as she started in math, got excited about applied math, and decided to study fluid mechanics. After getting her PhD at Stanford, she became a professor at the University of Auckland in New Zealand and then returned to Stanford where she has been a professor for 20 years. During this time, she has been an accomplished researcher, professor, mentor, and leader in the School of Earth, Energy & Environmental Sciences, the Institute for Computational & Mathematical Engineering (ICME), and Women in Data Science (WiDS).

When asked how she managed to juggle all of these things, Margot said she learned to not worry about making mistakes or striving for perfection, saying, “80% is perfect”, adding “I always felt I can't have it all. So you make choices, and there's always something that's got to give.” Cindy agreed that the busier she is, the better she manages her time, and when you have many balls in the air, often what you learn in one area can help you solve problems in another. 

In discussing the “imposter syndrome”, Margot said she had often felt like an imposter, and soon discovered this was a common feeling among students and faculty at Stanford. And it’s even stronger when you stand out, like a woman in STEM. It puts an extra burden on you to succeed to set the example for those who come after you. 

The pace of research in AI and deep learning contributes to feeling like an imposter. People publish very quickly and it's hard to understand what really good solid research is and what is just an idea. It gives people this sense that they're not on top. They forget the purpose of school is creating a lifelong interest in learning. “There's a lot of failure on the way to success. My favorite definition of an expert is somebody who's made every possible mistake.”

RELATED LINKS
Connect with Cindy Orozco on LinkedIN 
Find out more about Cerebras Systems
Connect with Margot Gerritsen on Twitter (@margootjeg) and LinkedIn
Find out more about Margot on her Stanford Profile

When Tahu Kukutai’s father went to school, he wasn’t allowed to speak his native language. If he did, he would be hit by his teachers. While the situation for the Maori people in Aotearoa (the Maori name for New Zealand) has improved somewhat since then, Tahu has dedicated her career to advocating for the rights of Indigenous peoples to preserve their native language, identity, communities, and culture. 

In today’s world, power over data is a central component of indigenous self-determination. Government agencies decide what data gets collected on whom and how it gets used, shared, and stored. There's a long history of data colonialism and state surveillance of indigenous communities designed for government priorities. Indigenous data sovereignty provides a framework to determine what data are collected and how it’s used, the ethical framework and governance, and the intended beneficiaries. 

Indigenous peoples have a different perspective on data. Most western frameworks about data protection, rights, and privacy are focused on the individual. The indigenous concept of data, data sovereignty and data rights are instead focused on the collective. Indigenous data are any information that impacts both the individual and the collective such as natural resources, nations, peoples, traditional and cultural information, oral histories, and ancestral knowledge. Tahu is advocating for indigenous data sovereignty as a founding member of the Māori Data Sovereignty Network Te Mana Raraunga and the Global Indigenous Data Alliance (GIDA) that works to control the collection, ownership, and application of data about their people, territories, and natural resources.

While there's a growing willingness from government agencies to engage in conversations around Maori data sovereignty, she says there is a lot of work ahead. Private sector big data organizations also have a major role to play, and she hopes they will be more proactive about doing this work internally to be relevant to the growing Maori population. Maori currently make up 16-18% of the Aotearoa population, compared to roughly 5% indigenous populations in the US, Canada, and Australia. 

Tahu says that advocating for any issue fundamental to the wellbeing of the Maori, whether it’s language revitalization, land rights, or water rights, has always been driven by the Maori. “I have huge optimism that change can occur, but it will only ever be driven by ourselves. That doesn't mean that others don't have a role to play and there are some very positive steps,” she says. “There's incredible resilience, fortitude, and tenacity in indigenous communities around the world who have refused to let go of their identity. The challenge is to move from surviving to thriving.”

RELATED LINKS
Connect with Tahu on LinkedIN and Twitter
Find out more about National Institute of Demographic and Economic Analysis (NIDEA)
Find out more about University of Waikato 
Connect with Margot Gerritsen on Twitter (@margootjeg) and LinkedIn
Find out more about Margot on her Stanford Profile

Allison Koenecke, who received her PhD from Stanford’s Institute for Computational and Mathematical Engineering (ICME), describes how her experiences in academia and industry shaped her decision to return to academia. Currently a postdoc at Microsoft Research in the Machine Learning and Statistics group, she starts as an Assistant Professor of Information Science at Cornell University next year. Her research interests lie at the intersection of economics and computer science, with projects focusing on fairness in algorithmic systems and causal inference in public health.
​
Allison says in her career so far, she has always tried to keep as many doors open as possible but recognized, at some point, you have to start closing doors and specialize. After getting her bachelor’s degree in mathematics from MIT, she worked in economic consulting for a few years and realized she wanted to do something with more social benefit. While she was working in industry and during summer internships, she kept in touch with professors and kept up with her research so she could have that option open if she wanted to go back to school. 

One of the main reasons she chose to stay in academia was industry and government did not offer what she was looking for. For example, if you stay in industry long-term and your research is critiquing big tech companies, you may run into roadblocks or not be seen as a neutral third-party observer, as you would be seen in academia. Or at a government think tank, your work wouldn't necessarily have as much reach as in academia. But even more, a lot of the reason she stayed in academia was the people. 

Allison’s research is interdisciplinary and falls into two categories. The first is a fairness in online services and algorithmic services, such as speech-to-text or online ads and looking at the racial disparities in those services. And the second branch is on causal inference, which is usually applied to things like public health. Most of her thesis focuses on fairness with the services that we use every day.

One of her research projects is about Google ads used to enroll people in food stamps and how to make decisions about fairness when it costs more to show those ads to Spanish speakers versus English speakers. She is also doing fairness research on racial disparities on speech-to-text systems developed by large tech companies to ensure systems are usable for African American populations that may not able to use their tools simply because they speak with a different variety of English than standard English. She says you need to have people thinking about fairness problems at all steps of the pipeline before you build a product that might harm certain groups of people. She’s hoping to bring awareness to different blind spots to make sure technology actually works for everyone.

RELATED LINKS
Connect with Allison on LinkedIN and Twitter
Find out more about the Microsoft Research Machine Learning and Statistics group
Find out more about Cornell University Information Science
Connect with Margot Gerritsen on Twitter (@margootjeg) and LinkedIn
Find out more about Margot on her Stanford Profile

Though Karina showed an early aptitude in math, her high school counselor advised her against pursuing an engineering degree. She ignored his advice and went on to earn her undergrad degree in mechanical engineering from the University of Rhode Island and a PhD in Aeronautics from Caltech. She landed her first job as a speech-to-text engineer at TRW where she was awarded her first patent. She then moved on to technology transfer as a patent agent at the Jet Propulsion Lab. She bounced back to academia, managing corporate partnerships for Caltech, and then returned to industry as Google’s University Lead for Google Cloud AI/Machine Learning. She is now at SAP as Global Head of Academies and University Alliances, continuing to connect industry and academia.

In her diverse career spanning business and education, she has seen increasing power concentrated in big tech companies through their ownership of immense datasets and computational power. Companies are also attracting talent away from universities that are now having a hard time hiring enough computer science faculty. She says there are some creative ways to bring back some balance by companies hosting visiting faculty and industry partners coming in to teach at universities.

Karina is also very concerned about ensuring fairness in data science. She explains that it’s not just the software that's being developed, but the datasets that are used to create predictive models. If a company just collects data from one demographic and then applies it to everyone, that introduces bias, and then the algorithms amplify these biases. She believes that the only way to address this is to have more ethnic, gender and geographic diversity in the field of data science. She sees a vital need to encourage more women and minorities to enter the field to bring diverse perspectives to data science.

For people interested in pursuing a career in data science, she advises gaining the basic skills in math, science, and programming languages, but the most important quality is the ability to learn because everything is constantly changing. She recommends keeping your options open, acquiring as many skills as possible, and sharpening your interpersonal skills. Karina also says to challenge yourself. “We don't grow in a space of comfort. You grow when you're challenged, it's okay to be uncomfortable because that's likely the place of greatest growth. There’s no such thing as failure, you either win or you learn.”  

RELATED LINKS
Connect with Karina on LinkedIn and Twitter
Find out more about SAP
Connect with Margot Gerritsen on Twitter (@margootjeg) and LinkedIn
Find out more about Margot on her Stanford Profile

Fatima Abu Salem grew up in Lebanon and has focused her data science research on addressing critical challenges in the region, including problems around the Syrian refugee crisis, the water quality in Lebanon and their irrigation requirements for farmers. 

Fatima explains that her life journey is really enshrined in the conflict of the region. She was born to a Lebanese mother and a Palestinian father who had refugee status so she inherited this status. This meant there were many restrictions on what career she could pursue and her dream to become a doctor was out of reach. But she was good in math so majored in mathematics at the American University of Beirut and one of the only jobs available to her as a Palestinian was able to become a high school teacher. She taught for a few years but realized it wasn’t what she wanted to do with her life. 

At the time, she was working on her master’s thesis in Algorithmic Number Theory at the American University in Beirut and decided she wanted to pursue a PhD. She emailed Oxford’s Computing Lab and a prominent professor there, Richard Brent, wrote back and invited her come study there. While at Oxford, she studied computer science, building on her knowledge of mathematics. After Oxford, she returned to AUB in Lebanon in the fall of 2004 as a faculty member.Data science was gathering momentum, so she started learning about it and was especially excited about applying data science for the public good. 

Fatima is currently using data science to help out with some of the most urgent and critical questions in Lebanon and in the Middle East. She focused on what mattered to her most about the livelihood of people such as the plight of refugees, fake news, the demand on the power grid and infrastructure, food security and healthcare. She wanted to help empower the Lebanese institutions to learn how to make better predictions so they can be more prepared than when they were taken by surprise by a million and a half refugees coming to Lebanon.

She has built collaborations with experts from the medical, agriculture, and public health schools for different research projects for social impact. One is trying to predict an irrigation metric to help farmers optimize their water usage so they can increase crop yields and reduce power consumption. Another project is trying to predict dementia with a public health school and a third project trying to predict birth defects using data using air pollution data.

In Fatima’s talk at the 2021 WiDS Conference, Doing Data Science in Data Deserts, she discusses these projects and others such as quantifying anti-refugee bias across Lebanese news corporations, fake news detection, predicting primary health care demand by Syrian refugees in Lebanon, and understanding Syrian refugee mobility in Turkey. 

RELATED LINKS
Connect with Fatima on LinkedIn and Twitter
Find out more about American University of Beirut (AUB)
Connect with Margot Gerritsen on Twitter (@margootjeg) and LinkedIn
Find out more about Margot on her Stanford Profile

Louvere Walker-Hannon has worked at MathWorks (the company that makes MATLAB) for over 21 years, where she’s also a STEM Ambassador. She studied biomedical engineering as an undergraduate at Boston University and did graduate work at Northeastern University in geographic information technology with a specialization in remote sensing.She loved working with MATLAB as an undergraduate and when MathWorks came to the career fair when she graduated, she sought them out, got an interview, and has been working there ever since.

She says there are both technical and non-technical skills that are valuable in the field of data science. Technical skills include coding, some programming, a foundation in mathematics, some statistics, and in some cases physics. Non-technical skills are also very important. It’s critical to be able to communicate your findings clearly using a variety of techniques. She says stay away from technical jargon and communicate as if you’re having a conversation. A second important skill is active listening, to be open to suggestions from others, especially those who are new to the field.

She explains that there are also barriers to people entering the field of data science. For some people, coding is a barrier to engagement with data science as many people in STEM professions are not comfortable with coding. MathWorks is doing more development to provide user interfaces or apps to give people a starting point without having to rely on writing code.

There are also concerns about model interpretability where it’s difficult to get insights into how certain models work. More people are gaining awareness about the topic, and that's leading them to explore how to implement it and ask why it’s important. She is noticing that more people are trying to incorporate model interpretability into their data science applications.

One of the systemic barriers is implicit bias. People are used to working with and being around people with certain characteristics. And in a work setting, there could be a project coming up, and there are several individuals who could work on this project. Many times, the people selected to work on a project tend to be the same individuals. But then it begs the question, when do others get the chance? 

There’s still a lot of under-representation from various population groups in data science. Even if people from an under-represented group have the skills and education, if they don’t feel like they belong, that can impact their productivity. It’s important to build a sense of community and have someone guide the person, make them feel welcome and help them become a part of the culture, so they can understand what they can do in order to thrive. 

Louvere is also a STEM ambassador at MathWorks where she volunteers in STEM advocacy and outreach in schools or on STEM panels. She loves hearing high school age and younger students at science fairs talk about their projects and see how proud they are of their work. This gives her hope that young people are excited about data, about analysis, and communicating their insights to others. 

RELATED LINKS
Connect with Louvere on LinkedIn and Twitter
Find out more about Mathworks
Connect with Margot Gerritsen on Twitter (@margootjeg) and LinkedIn
Find out more about Margot on her Stanford Profile