 Okay, good morning everyone and welcome. So we're here at the first of three lunches to celebrate our new associate professors. And this morning we'll be celebrating Associate Professor Brett Savoy in Chemical Engineering, Associate Professor Jian Jin in Ag and Biological Engineering and Associate Professor Tim Rogers in Electrical and Computer Engineering. So we're looking forward to hearing a little bit from each of you, reflecting on your past to the successes that we're celebrating today and maybe your thoughts about where you're heading in the future. So as a land grant institution, we take our mission seriously to provide a high quality education to as many students as we can. But we also have a goal and aim to be at the pinnacle of excellence at scale, to be at the very best and second to none. And some will say you can't be both. You can be very large and maybe pretty good, maybe even very good, but you can't be second to none in excellence. So we've set that very ambitious goal for ourselves and I'm proud that we've done that and I don't underestimate how challenging that is. But we've made a lot of progress and the next step will be a big one. So for those of you that are after a promotion, it's a good time to sort of pause and stop and think like what's next. There are several different pathways to the pinnacle and choose the one that's best for you, but don't settle for being anything less than the very best in your field. That's what we'd like to help you all achieve. So with that, I will turn the mic over to Sangtae Kim, head of chemical engineering to introduce our first speaker. Thank you, Mark. And I'll keep my introduction very brief because I know Brett has an amazing set of materials that he wants to share with you. Some say that chemical engineering is the branch of engineering that rests on the scientific disciplines of chemistry and physics. So not surprisingly, our honoree today has a double major in physics and chemistry from Texas A&M followed by a PhD in computational chemistry from Northwestern and then postdoctoral studies at Caltech and joined our faculty in 2017. Since joining our faculty, Brett has shown great leadership and vision at the intersection between data science and the opportunities in chemical engineering. So with that, Brett, the floor is yours. All right, thank you, Sang. It's a tremendous honor to have received tenure, especially at this institution. So yeah, this is a wonderful event to sort of inaugurate that and give me a chance to reflect. I'm really appreciative of being here. So this is a very hard talk to give. I have 10 minutes to summarize how I got here and gosh, I give talks all the time and this was a really difficult one for me to compose. So at the heart of this talk is a lot of thank yous. That's really the main gist with a little bit of time reserved for my sort of ideas on the future. Okay, so I was born in Flint, Michigan to Dennis and Teresa Savoy and the first thank you goes out to them. So I've been extraordinarily privileged to have a wonderfully supportive family, not only my immediate family, so my father was in the auto industry, that's the origin of us being in Flint and my extended family throughout the region to this day. I have very fond memories of that and I still go and visit the area. So thank you to Flint, Michigan, my immediate family, we moved to Texas when I was very young. I was reared there and that's why when I came college age, I went to another institution, not Purdue, I went to Texas A&M. And there's another thank you here at Texas A&M. I'd never thought about being an academic and I never thought about having an advanced degree. It didn't come from an academic background and there was a physics professor named Dave Tobak who picked me out of this very large class and he essentially told me, you should really think about doing this for your career. A class of 150 people, there'd been a challenge exam and I'd done very well on this challenge exam. So he set up a meeting with me proactively and I would never be a professor today if that action hadn't been taken, genuinely. So that's how I ended up becoming a physics major. And then the second thank you is to Marcetta Dernsberg. So I was studying physics and chemistry because my parents didn't really know what do you do with a physics degree, they wanted me to be a doctor. So I was taking all these chemistry courses to be, keep that open as an option. I was picking up the second degree and my research in chemistry, I just fell in love with it. The feedback I could get, how quickly the feedback was compared to physics. So I went to, got my PhD in chemistry. That's how I ended up at Northwestern. At Northwestern, I was really spoiled for outstanding mentors. So Mark Ratner and Tobin Marks were just phenomenal world-class scientists, also amazing people. So for my PhD, I ended up in Northwestern, Northwestern. And I was spoiled. I had world-class scientists that were my mentors. Even though I was a theoretical, purely theoretical PhD student, Tobin, who's an experimentalist, had a great history of having theoretical people in his group. So he just kind of wanted, he liked having me around, so to speak. And Mark Ratner was the theoretical component of my mentorship and they're really astonishing people. I think a lot of you actually turns out, Tobin has a very strong Purdue connection in chemical engineering research. So he doesn't actually collaborate on any of the projects I'm currently on, but he still is a collaborator to this day. And Mark also has a Purdue connection and they're really phenomenal people and I owe them a tremendous amount. So then I went over to Caltech, where I worked with actually also a former Aggie. So Tom Miller was a theoretical chemist at Caltech, sort of one of the rising stars in the field. And I did my postdoc there. And Tom is really the springboard that provided my ultimate opportunity to end up during the faculty search, opening up and ending up at Purdue. So that's my physical journey. That's my physical journey. Now along that journey, especially I wanna start at Purdue, I'm gonna go in chronological order. There's a lot of additional thank yous. So the first person I really interacted with was saying, so I submitted my application within a week or two. I got a call from saying, saying you wanted to, interview me, we talked a little bit. And also when I arrived, we had dinner the first night. So my first image of Purdue was really about saying. So here's a couple lessons, saying shared with me. Oops. First one is, from day one, you aren't a guest here, you are a partner. I'm sort of paraphrasing here. So saying really lives us out. So as a young faculty member, I always felt heard at faculty meetings. You would always insist on hearing from the younger faculty. We had standing monthly meetings, talking about just what was going on. And yeah, I really learned a lot by the way that he has been head of the department and the kind of culture that he cultivated. He's also taught me the difference between strategy and tactics. So if you get this job, you're probably pretty good tactically in terms of how you approach a specific problem in your area. But strategic thinking is something that you don't usually come in with as an assistant professor. And I think saying is a brilliant strategist and he's taught me a lot along those lines. Second person, chronologically speaking, Brian Beduris. Where's Brian at? Oh, there you are. Brian Beduris, he's tiny and he's in the back. Brian is the second person I want to call out who's very influential, starting on my interview. So there's a quote here that is not from Brian, okay? So during, not all of my interviews were as good as my Purdue interview. Shockingly, they weren't all as good as this one. One of the interviews I was on, I remember a faculty member saying to me, I mean, is this really the future of chemical sciences? All this applied stuff? Like, and it wasn't meant in a complimentary fashion. And in contrast, I remember Brian telling me, he probably just tells us to everyone. He says, your application package has the most vision. It's just, you can see where you're going. It's so well put together. I mean, I was just so impressed by it. And he told me that on the interview and it really made me feel good. I don't know if Brian realizes that, but my impression of Purdue was strongly affected by Brian. Of course, after I got here, Brian taught me a lot on the technical side. If I could summarize it in one sentence, it would be that non-conjugated materials can conduct too. So I've been completely swayed towards these radical-based conductors. They're non-conjugated and they conduct. It's been a lot of fun to study that. And I think Brian is kind of astonishing because he embodies this concept of effective service. So I've never seen anyone, he's not just calling it in. He's always so effective on every service, opportunity that he has. That's really been an inspiration to me. Jim Crothers, I didn't meet Jim actually until after I arrived on campus and he's my faculty mentor. And Jim always told me, don't think about getting tenure. This is, and I think that this is wise because it's kind of like the economic principle that once a measure becomes a target, stops being a good measure. The thing that got you made me successful in the past. It wasn't agonizing about tenure. And so I appreciated that coming from my faculty mentor. He also told me to learn how to say no. For those of you who know Jim, he's good at this. He's good. And I think the corollary is very important which is learning how to say yes to the right things. It's a really important skill because obviously you can't say yes to everything so you have to learn how to say no. And I want to emphasize horizontal mentorship. So another one of the associate professors arrived on campus with me and I think there's horizontal mentorship which came in the form of Letty and Doe in my department. He, it's not easy being an assistant professor and sometimes it's really important to have other people who are telling you, yeah, it's not easy. When your grants aren't hitting and you're struggling with, the research isn't always working out perfectly. And so Letty and I really owe you a lot in the way of horizontal mentorship. And he also taught me that's okay to research perovskites. So I was really uncomfortable with perovskites. I didn't want to work on them. I was a Palmer person but Letty and proved to me that it's okay. And I want to say thank you to this sort of non-exclusive list of other mentors. So I have various sort of cute names for all of them. So like Jeff Greeley is my fellow computationalist. Steve and Jeff have taught me a ton about energetic, something I never knew anything about. And I can't, I'm unfortunately, I can't say thank you to everyone on here. The other aspect of mentorship is how I'm trying to, I'm trying to mentor people. So I'm extremely humbled by my research group, many of whom are here today. I remember to share one small anecdote. I can remember arriving here at Purdue in August and the first set of graduate students came to campus. And I can remember being nervous because I didn't think anyone was going to want to join my group. I was really, I said, you know, who's going to want to join the group of the new guy? You know, are they even going to set up meetings for me with me? And Aditi and Nick were my first two graduate students and Steven was my first postdoc. And they've both graduated now at this point. And they've been world-class students and they set the pattern, which hasn't been broken yet. I'm looking at all of you of having really astonishingly good graduate students. And I just want to say that I really believe in our obligation to, for mentorship, that, you know, my job isn't just to keep the lights on, it is to provide active mentorship. And without you all, I wouldn't have this job. And that extends to our undergraduates, certainly. I take this very seriously. I would not have this job. I might have a position in a research lab or something, but I wouldn't have this job if it wasn't for you all. So I'm going to very briefly just talk about two problems looking forward. So if you're interested in my research today, you can look at my website. We've published lots of papers. But I think 10 years of time for maybe picking some ambitious new targets. So two problems that I'm very interested in. So reaction prediction. So for a long time, we've worked really hard on, as computational modelers, we've worked hard on modeling a specific transition state, trying to find the mechanism of a given reaction where you know the reactants and you know the product. But there's a lot of problems where you don't know the product, reaction exploration. This is sort of a holy grail of you only know the reactant side of the arrow, can you predict the outcome of the reaction, only knowing that information. And if you could do this ab initio on a computer, you could discover new reactions, you could discover new catalysts. And I think we have some great ideas about how to do that. Another place where it's coming up is how materials degrade, which has for a long time been a sore spot for us because we can't predict that very well, how they degrade, we can predict other functional properties. So A goes to the question mark problem is one that I hope to make my name as an associate professor. Another problem is chemical deduction. So I do a lot on the machine learning side of things. And if I could summarize all of the machine learning research that's been going on in the past sort of five years, I would say that we have convincingly solved the supervised learning problem when it comes to molecular properties. There's plenty of caveats in terms of, the data intensity of this, but problem after problem after problem, we've shown that we're able to predict pretty much arbitrary molecular properties provided sufficient training data and provided a suitable model architecture. Now, but deduction problems are much more compelling because deduction problems are where we use existing knowledge to hypothesize an outcome that's under determined. So you don't have, intrinsically, you don't have enough data to conclusively have one and only one answer. And so we try to deduce based on what we know about reactivity relationships or otherwise chemical or chemical understanding. So I think the deduction problem is very exciting. You will be hearing more about that from me in over the coming years. All right, I will finish on this note. So this is celebrating my accomplishments. By far my biggest accomplishment is my family. So you all don't know this about me. A lot of you don't know this about me, but there's over eight billion people in the world and I found the best one and I convinced her to marry me. And we have beautiful children together. This is the rare photo. This is a selfie from just a couple of months ago. We're in the middle of the woods. We'd actually hiked out a couple of miles and cause always one of us is missing from the photo. So this is all of us in there. So thank you very much for that. I'm happy to answer some questions. And we probably have to hand the mic to whoever wants to ask a question because it's being a questions. Oh, okay. It's one more here, right? Oh, okay, yeah. Is it? In the photo is one missing, right? No. Because you said he won six, right? No, oh, maybe, maybe. Right now there's only five children. We only have five. We'll see about six. So I'm interested in perovskites for solar cell applications. Sure. What are the challenges, the current challenges there? Yeah. So the most important challenge for perovskites is stability, I would argue. The efficiencies have risen dramatically and we still haven't made proportional gains in stability. And so that's actually the focus of my work with Letian is we think we have ideas about how to stabilize these things, particularly the interfaces. So if you can get the bulk crystal seems to be fine if you can get the interfaces stabilized. And so that's where we're actually trying to do. You also have more design flexibility at the interfaces. So the bulk has very, you're very constrained on the perovskite constitution but at the surface you can functionalize it with the organics. And so there's exciting design problems and then it also interfaces with if you could stabilize these then that would have a quality of change and translating them into technology. Thanks for the question. Brett, we've talked about the deduction problem in some detail, but it might help because it dawned upon me what solving that problem might accomplish. Sure. Only after we had our discussions on this may be useful to just share with the group about well if you crack that, what follows? Yeah. What are the corollaries to that? Okay, absolutely, yeah. So I'm only too happy to hold forth on why I think this is important. So I'll give you two examples of deduction problems. So one is a reaction outcome. So if you put in, if you know what you put into a reaction you don't always get the same thing because sometimes people don't actually follow, there's a bunch of uncertainty about like what solvent did you use? Did you have these other additives? Was there a catalyst present or not? What temperature did you use? Just stating what was in the beaker doesn't actually uniquely determine the product. You might have starting material. You might have had side products. So deduction, the deduction problem is what a chemist does, an analytic chemist does when they look at spectra and they have to deduce the specific product that came out from a particular test reaction. Now if you were trying to do high throughput automated reaction screening which you can do today, the slow point is expert interpretation of those reaction outcomes. So it's great. You can test 10,000 different reaction conditions but unless you have a really low cost way of figuring out if you found, oh did I form that carbon-carbon bond, you're limited by expert deduction of which one of these was a useful outcome. And so a machine learning model that could actually do deduction would solve a problem like that. Another one that's really close to my heart is polymer. So a reason why soft materials have had less impact on machine learning or machine learning has had less impact on soft materials is because they're so processing dependent. If you just give me the constitution of a soft material, the actual properties of that material are so sensitive in how it was processed. So the constitution doesn't tell you its performance. That's also where deduction can come in. So you know the chemical formula, you maybe have a small number of experimental inputs, you can deduce the properties then in a way that you can't with an ordinary supervised learning method. Thank you. Okay, well let's thank Brett again. And thank you for coming back.