 Good afternoon, everyone, and thank you so much for joining us. My name is Savannah Peterson, joined by my co-host, Paul, for the afternoon. Very excited. Hello, Savannah. Hello, I'm pumped for this. It's our first bit together. Exactly, it's going to be fun. Yeah. We have a great guest to kick off with. We absolutely do. We're at Supercomputing 2022 today. And very excited to talk to our next guest. We're going to be talking about data at scale and data that really matters to us. Joining us, Kelly Gather. Thank you so much for being here. And you are with TAC. Tell everyone what TAC is. TAC is the Texas Advanced Computing Center at the University of Texas at Austin. And thank you so much for having me here. It is wonderful to have you. Your smiles contagious. And one of the themes that's come up a lot with all of our guests, and we just talked about it, is how good it is to be back in person, how good it is to be around our hardware community. At TAC, you did some very interesting research during the pandemic. Can you tell us about that? I can. I did. So when we realized, sort of mid-March, we realized that this was really not normal times. And the pandemic was going to touch everyone. I think a lot of us at the center and me personally, we dropped everything to plug in. And that's what we do. So UT's tagline is what starts here, changes the world. And TAC's tagline is Powering Discoveries that change the world. So we're all about impact. But I plugged in with a research group there at UT Austin, Dr. Lauren Myers, who's an epidemiologist. And just we figured out how to plug in and compute so that we could predict the spread of COVID-19. And you did that through the use of mobility data, cell phone signals. Tell us more about what exactly you were choreographing. Yeah, so that was really interesting. SafeGraph, during the pandemic, made their mobility data. Typically it was used for marketing purposes to know who was going into Walmart. You know, fences for advertising. Yeah, absolutely. They made all of their mobility data available for free to people who were doing research and plugging in trying to understand COVID-19. I picked that data up and we used it as a proxy for human behavior. So we knew we had some idea. We got weekly mobility updates, but it was really mobility all day long. You know, anonymized. I didn't know who they were. By cell phones across the US. By census plot group or zip code if we wanted to look at it that way. And we could see how people were moving around. We knew what their neighbor, their home neighborhoods were. We knew how they were traveling or not traveling. We knew where people were congregating and we could get some idea of how people were behaving. Were they really locking down or were they moving in their neighborhoods or were they going outside of their neighborhoods? What a fascinating window into our pandemic lives. So now that you were able to do this for this pandemic, as we look forward, what have you learned? How quickly could we forecast? What's the prognosis? Yeah, so we learned a tremendous amount. I think during the pandemic, we were reacting. We were really trying. It was an interesting time as a scientist. We were reacting to things almost as if the earth was moving underneath us every single day. So it was something new every day. And I've told people since I haven't worked that hard since I was a graduate student. So it was really daylight to dark 24 seven for a long period of time because it was so important. And we knew we were being a part of history and affecting something that was going to make a difference for a really long time. And I think what we've learned is that indeed there is a lot of data being collected that we can use for good. We can really understand if we get organized and we get set up, we can use this data as a means of perhaps predicting our next pandemic or our next outbreak of whatever it is. Almost like using it as a canary in the coal mine. There's a lot in human behavior we can use. Given all the politicization of this last pandemic, knowing what we know now, making us better prepared in theory for the next one. How confident are you that at least in the US we will respond proactively and effectively when the next one comes around? Yeah, I mean that's a great question. And I certainly understand why you ask. I think in my experience as a scientist, certainly at TAC, the more transparent you are with what you do and the more you explain things. Again, during the pandemic, things were shifting so rapidly, we were reacting and doing the best that we could. And I think one thing we did right was we admitted where we felt uncertain. And that's important. You have to really be transparent to the general public. I don't know how well people are going to react. I think if we have time to prepare, to communicate, and always be really transparent about it, I think those are three factors that go in to really increasing people's trust. I think you nailed it. And especially during times of chaos and disaster, you don't know who to trust or what to believe. And it sounds like providing a transparent source of truth is so critical. How do you protect the sensitive data that you're working with? I know it's a top priority for you and the team. It is, it is. And we've adopted the medical mantra, do no harm. So we feel a great responsibility there. There's two things that you have to really keep in mind when you've got sensitive data. One is the physical protection of it. And so that's governed by federal rules, HIPAA, FERPA, whatever kind of data that you have. So we certainly focus on the physical protection of it. But there's also sort of the ethical protection of it. What is the quote? There's lies, damn lies and statistics. Yes, twain. Yeah, so you really have to be responsible with what you're doing with the data, how you're portraying the results. And again, I think it comes back to transparency. Is basically, if people are going to reproduce what I did, I have to be really transparent with what I did. Yeah, I think that's super important. And one of the themes with HPC that we've been talking about a lot too is do people trust AI? Do they trust all the data that's going into these systems? And I love that you just talked about the storytelling aspect of that because there is a duty. You can cut data kind of however you want. I mean, come from marketing background and we can massage it to do whatever we want. So in addition to being the deputy director at TAC, you are also the DEI officer and diversity, I know, is important to you, probably both as an individual but also in the work that you're doing. Talk to us about that. Yeah, I mean, I'm very passionate about diversity, equity and inclusion in a sense of belongingness. I think that's one of the key aspects of it. I got a computer science degree back in the 80s. I was akin to a unicorn in an engineering computer science department. But I was really lucky in a couple of respects. I had a father that was into science that told me I could do anything I wanted to set my mind to do. So that was my whole life was really having that support system. He was cheers to dad. Yeah, oh yeah. And my mom as well. Actually, they were educators. I grew up in that respect, very, very privileged but it was still really hard to make it. And I couldn't have told you back in that time why I made it and others didn't, why they dropped out. But I made it a mission probably back, gosh, maybe 10, 15 years ago that I was really going to do all that I could to change the needle. And it turns out that there are a number of things that you can do, grassroots. There are certainly best practices. There are rules. And there are things that you really, best practices to follow, to make people feel more included in an organization, to feel like they belong, it shared mission. But there are also clever things that you can do with programming to really engage students, to meet people and students where they are interested and where they are engaged. And I think that's what we've done over the course of our programming, over the course of about maybe since 2016, we have built a lot of programming at TAC that really focuses on that as well. Because I'm determined the needle is going to change before it's all said and done. It just really has to. So what progress have you made and what goals have you set in this area? Yeah, that's a great question. So, you know, at first I was a little bit reluctant to set concrete goals because I really didn't know what we could accomplish. I really wasn't sure what grassroots efforts was going to be able to do. You're so honest. You can tell how transparent you are with the data as well. That's great. Yeah, I mean, if I'm really, most of the successful work that I've done is both a scientist and in the education and outreach space is really trust relationships. If I break that trust, I'm done. I'm no longer effective. So yeah, I am really transparent about it. But what we did was, you know, the first thing we did was we counted, you know, to the extent that we could, what does the current picture look like? Let's be honest about it. Start where we are. It was not a pretty picture. I mean, we knew that anecdotally. It was not going to be a great picture, but we put it out there and we leaned into it. We said, this is what it is. We, you know, I hesitated to say, we're going to look 10% better next year because I'm going to be honest, I don't always know. We're going to do our best. The things that I think we did really well was that we stopped to take time to talk and find out what people were interested in. It's almost like being present and listening. My grandmother had a saying, you have two ears and one mouth for a reason. Just respect the ratio. Oh, I love that. Yeah, and I think it's just been building relationships, building trust, really focusing on making a difference, making it a priority. And I think now what we're doing is we've been successful in pockets of people in the center and we are getting everybody on board. There's something everyone can do. But the problem you're addressing doesn't begin in college, it begins much, much earlier. And there's been a lot of talk about STEM education, particularly for girls, how they're pushed out of the system early on, also for people of color. Do you see meaningful progress being made there now after years of lip service? I do, I do, but it is, again, grassroots. We do have a researcher who is a former teacher at the center, Carol Fletcher, who is doing research and CS for all. We know that the workforce, so if you work from the current workforce or projected workforce backwards, we know that digital skills of some kind are going to be needed. We also know we have a shortage. There's debate on how large that shortage is, but roughly about one million unmet jobs was projected in 2020, it hasn't gotten a lot better. We can work that problem backwards. So what we do there is a little like a scatter shot approach. We know that people come in all forms, all shapes, all sizes, they get interested for all different kinds of reasons. We expanded our set of pathways so that we can get them where they can get on to the path. All the way back K through 12, that's Carol's work. Rosie Gomez at the center is doing sort of the undergraduate space. We've got Don Hunter that does it, middle school, high school space. So we are working all parts of the problem. I am pretty passionate about what we consider opportunity youth, people who never had the opportunity to go to college. Is there a way that we can skill them and get them engaged in some aspect and perhaps get them into this workforce? I love that you're starting off so young. So give us an example of one of those programs. What are you talking to kindergartners about when it comes to CS education? You know, I mean gaming. Right, it's what everybody can wrap their head around. So most kids have had some sort of gaming device. You talk in the context of something they understand. I'm not going to talk to them about high performance computing. It would go right over their heads. And I think, you know, I'll go back to something that you said Paul about girls were pushed out. I don't know that girls are being pushed out. I think girls aren't interested in things that are being presented. And I think they leave. I think you're generous. Yeah, I mean I was a young girl and I don't know why I stay. Well, I do know why I stayed with it because I had a father that saw something in me and I had people at critical points in my life that saw something in me that I didn't see. But I think if we change the way we teach it, maybe in your words, they don't get pushed out or they won't lose interest. There's some sort of computing in everything we do. Absolutely. There's also the bro culture which begins at a very early age. Yeah, that's a different problem. That's just having boys in the classroom. Absolutely, you got it. That's a whole other thing. Last question for you, when we are sitting here, well actually, it's two-parter, let's put it that way. Is there a tool or something you wish you could flick a magic wand that would make your job easier? Can you identify the linchpin in the DEI challenge or is it all still prototyping and iterating to figure out the best fit? Yeah, that's a wonderful question. I can tell you what I get frustrated with, is that... That counts. Is that I feel like a lot of people don't fully understand the level of effort and engagement it takes to do something meaningful. The commitment to a program. The commitment to a program, and there is no one and done. No. And in fact, if I do that, I will lose them forever. They will be lost in this space forever. Rather, the engagement is really sort of time-intensive. It's relationship-intensive. But there's a lot of follow-up to you. And the amount of funding that goes into this space really is not equal to the amount of time and effort that it really takes. And I think what you work in this space, you realize that what you gain is really more of, it really feels good to make a difference in somebody's life, but it's really hard to do on a shoestring budget. So if I could kind of wave a magic wand, I would increase understanding. I would get people to understand that it's all of our responsibility. Everybody is needed to make the difference, and I would increase the funding that goes to the programs. I think that's awesome. Kelly, thank you for that. You all heard that. More funding for diversity, equity, and inclusion. Please, Paul, thank you for a fantastic interview. Kelly, hopefully everyone is now inspired to check out TAC, perhaps become a longhorn, hook them, and come deal with some of the most important data that we have going through our systems and predicting the future of our pandemics. Ladies and gentlemen, thank you for joining us online. We are here in Dallas, Texas at Supercomputing. My name is Savannah Peterson, and I look forward to seeing you for our next segment.