 Hey everyone, welcome back to theCUBE's live coverage of Women in Data Science WIDS 2022, coming to you live from Stanford University. I'm Lisa Martin. My next guest is here, Nandi Leslie, Dr. Nandi Leslie, Senior Engineering Fellow at Raytheon Technologies. Nandi, it's great to have you on the program. Oh, it's my pleasure, thank you. This is your first WIDS you were saying before we went live. That's right. What's your take so far? I'm absolutely loving it. I love the camaraderie and the community of women in data science. You know, what more can you say? It's amazing. It is. It's amazing what they built since 2015, that this is now reaching 100,000 people, 200 online event. It's a hybrid event. Of course, here we are in person and the online event going on, but it's always an inspiring energy filled experience in my experience of WIDS. I'm thoroughly impressed at what the organizers have been able to accomplish. And it's amazing that you've been involved from the beginning. Talk to me. So you were a senior engineering fellow at Raytheon. Talk to me a little bit about your role there and what you're doing. Well, my role is really to think about our customers' most challenging problems, primarily at the intersection of data science and the intersectional fields of applied mathematics, machine learning, cybersecurity. And then we have a plethora of government clients and commercial clients. And so what their needs are beyond those sub fields as well, I address. And your background is mathematics. Yes. Have you always been a math fan? I have, I actually have loved math for many, many years. My dad is a mathematician and he introduced me to mathematical research and the sciences at a very early age. And so yeah, I went on. I studied a math degree at Howard undergrad and then I went on to do my PhD at Princeton in applied math and later did a postdoc in the math department at University of Maryland. And how long have you been with Raytheon? I've been with Raytheon about six years. Yeah, and before Raytheon, I worked at a small to mid-sized defense company, defense contracting company in the DC area, systems planning and analysis. And then prior to that, I taught in a math department where I also did my postdoc at University of Maryland College Park. You have a really interesting background. I was doing some reading on you and you have worked with the Navy. You've worked with very interesting organizations. Talk to the audience a little bit about your diverse background. Awesome, yeah. I've worked with the Navy on submarine force security and submarine tracking and localization, sensor performance, also with the Army and the Army Research Laboratory research at the intersection of machine learning and cyber security. Also looking at game theoretic and graph theoretic approaches to understand network resilience and robustness. I've also supported the Department of Homeland Security and other government agencies, other governments, NATO. Yeah, so I've really been excited by the diverse problems that our various customers have brought to us. Well, you get such great experience when you are able to work in different industries in different fields and that really just really probably helps you have such a much diverse kind of diversity of thought with what you're doing even now with Raytheon. Yeah, it definitely does. It helped me build like a portfolio of topics that I can address. And then when new problems emerge, then I can pull from a toolbox of capabilities and the solutions that have previously been developed to address those wide array of problems, but then also innovate new solutions based on those experiences. So I've been really blessed to have those experiences. Talk to me about some, one of the things I heard this morning in the session I was able to attend before we came to set was about mentors and sponsors. And I actually didn't know the difference between that until a few years ago, but it's so important. Talk to me about some of the mentors you've had along the way that really helped you find your voice in research and development. Definitely, I mean beyond just the mentorship of my family and my parents, I've had amazing opportunities to meet with wonderful people who've helped me navigate my career. One in particular I can think of is, and I'll name a number of folks, but Dr. Carlos Castillo Chavez was one of my earlier mentors. I was an undergrad at Howard University. He encouraged me to apply to his summer research program in mathematical and theoretical biology, which was then at Cornell. And he just really developed an enthusiasm with me for applied mathematics. And for how it can be, mathematics that is can be applied to epidemiological and theoretical immunological problems. And then I had an amazing mentor in my PhD advisor, Dr. Simon Levin at Princeton, who just continued to inspire me in how to leverage mathematical approaches and computational thinking for ecological conservation problems. And then since then I've had amazing mentors through just a variety of people that I've met through customers who've inspired me to write these papers that you mentioned in the beginning. You've written 55 different publications so far. 55 and counting, I'm sure, right? Well, I hope so. I hope to continue to contribute to the conversation in the community within research and specifically research that is computationally driven that really is applicable to problems that we face, whether it's cybersecurity or machine learning problems or others in data science. What are some of the things, you're giving a tech vision talk this afternoon. Talk to me a little bit about that and maybe the top three takeaways you want the audience to leave with. Yeah, so my talk is entitled Unsupervised Learning for Network Security or Network Intrusion Detection, I believe, and essentially three key areas I want to convey are the following, that unsupervised learning, that is the mathematical and statistical approach which tries to derive patterns from unlabeled data is a powerful one and one can still innovate new algorithms in this area. Secondly, that network security and specifically anomaly detection and anomaly-based methods can be really useful to discerning and ensuring that there is information confidentiality, availability and integrity in our data. C-I-A triad. There you go, you know. And so in addition to that, that there is this wealth of data that's out there, it's coming at us quickly. You know, there are millions of packets to represent communications and that data has, it's mixed in terms of there's categorical or qualitative data, text data, along with numerical data and it is streaming, right? And so we need methods that are efficient and that are capable of being deployed real-time in order to detect these anomalies which we hope are representative of malicious activities and so that we can therefore alert on them and thwart them. It's so interesting that the amount of data that's being generated and collected is growing exponentially. There's also some concerning challenges not just with respect to data that's reinforcing social biases but also with cyber warfare. I mean, that's a huge challenge right now. We've seen from a cybersecurity perspective in the last couple of years during the pandemic a massive explosion in anomalies in social engineering and companies in every industry have to be super vigilant and help the people understand how to interact with it, right? There's a human component. Oh, for sure. There's a huge human component. There are these phishing attacks that are really a huge source of the vulnerability that corporations, governments and universities face and so to be able to close that gap and the understanding that each individual plays in the vulnerability of a network is key and then also seeing the link between the network activities or the cyber realm and physical systems, right? And so these, especially in cyber warfare, a remote cyber attack, unauthorized network activities can have real implications for physical systems. They can stop a vehicle from running properly in an autonomous vehicle. They can impact a SCADA system that's there to provide HVAC, for example, and much more grievous implications. And so definitely there's the human component. Yes, and humans being so vulnerable to those social engineering that goes on in those phishing attacks and we've seen them get more and more personal, which is challenging. You're talking about sensitive data, personally identifiable data, using that against someone in cyber warfare is a huge challenge. Oh yeah, certainly. And it's one that computational thinking and mathematics can be leveraged to better understand and to predict those patterns and that's a very rich area for innovation. What would you say is the power of computational thinking in the industry? In industry at large? Yes, I think that it is such a benefit to a burgeoning scientist. If they want to get into industry, there are so many opportunities because computational thinking is needed. We need to be more objective and it provides that objectivity and it's so needed right now, especially with the emergence of data and across industries. So there are so many opportunities for data scientists, whether it's in aerospace and defense like Raytheon or in the health industry. And we saw with the pandemic the utility of mathematical modeling. Yes. There are just so many opportunities. Yeah, there's a lot of opportunities and that's one of the themes I think of Woods is just the opportunities, not just in data science and for women and there's obviously even high school girls that are here which is so nice to see those young, fresh faces but opportunities to build your own network and your own personal board of directors, your mentors, your sponsors. There's tremendous opportunity in data science and it's really all-encompassing. At least from my seat. Oh yeah, no, I completely agree with that, yeah. What are some of the things that you've heard at this Woods event that inspire you going, we're going in the right direction. If we think about International Women's Day tomorrow, breaking the bias is the theme. Do you think we're on our way to breaking that bias? Definitely, there was a panel today talking about the bias in data and in a variety of fields and how we are discovering that bias and creating solutions to address it. So there was that panel, there was another talk by a speaker from Pinterest who had presented some solutions that her and her team had derived to address bias there in image recognition and search. So I think that we've realized this bias and AI ethics, not only in these topics that I've mentioned, but also in the implications for getting a loan. So economic implications as well. And so we're realizing those issues and bias now and AI and we're addressing them. So I definitely am optimistic. I feel encouraged by the talks today at Woods that not only are we recognizing the issues, but we're creating solutions. Right, taking steps to remediate those so that ultimately going forward, we know it's not possible to have unbiased data. That's not humanly possible or probably mathematically possible, but the steps that they're taking, they're going in the right direction and a lot of it starts with awareness. Exactly. Of understanding there is bias in this data, regardless all the people that are interacting with it and touching it and transforming it and cleaning it, for example, that that's all influencing the veracity of it. Oh, for sure, exactly. And I think that there are for sure solutions that are being discussed here, papers written by some of the speakers here that are driving the solutions to the mitigation of this bias in data problem. So I agree 100% with you that awareness is, you know, half the battle, if not more. And then, you know, that drives creation of solutions. And that's what we need, the creation of solutions. Nandi, thank you so much for joining me today. It was a pleasure talking with you about what you're doing with Raytheon, what you've done in your past with mathematics and what excites you about data science going forward. We appreciate your insights. Thank you so much, it was my pleasure. Good. For Nandi Leslie, I'm Lisa Martin. You're watching theCUBE's coverage of Women in Data Science 2022. Stick around, I'll be right back with my next guest.