The Bridge Builders: Success Stories in Industry Collaboration (Part 2)

In this episode of Bridging the Gap, we are going to talk about how to build a bridge between the lab and the marketplace. We are excited to be joined by Dr. Stevie Chancellor, a University of Minnesota faculty member and professor in the College of Computer Science and Engineering. Dr. Chancellor has successfully navigated the art of the "win-win" partnership between academia and industry and has graciously agreed to share her advice, and her stories with us.

Hosted by University of Minnesota Corporate Research Partnerships Managers Leza Besemann and Jason Stockman, "Bridging the Gap" is a practical podcast series designed to help university faculty navigate the complexities of collaborating with the private sector. 

Transcript

Jason Stockman 0:09

Welcome back to Demystifying University and Industry Partnerships. I'm your host, Jason Stockman, a corporate research partnerships manager here at the University of Minnesota. On our last podcast, we were joined by Dr. Gerard Cramer, a University of Minnesota faculty member and professor in the College of Veterinary Medicine. Dr. Cramer shared his advice and his stories about his personal experiences with academic and industry research partnerships. And based on all the positive feedback we received with our interview with Dr. Kramer, we're going to talk to another University of Minnesota faculty member about how she built a bridge between the lab and the marketplace. I'm excited to be joined today by Dr. Stevie Chancellor, who's a professor in the Department of Computer Science and Engineering. Dr. Chancellor has also successfully navigated the art of the win-win partnership between academia and industry and has graciously agreed to share her advice and her stories with us today. So, Dr. Chancellor, thank you for being here with us today.

Dr. Stevie Chancellor 1:04

Thanks so much for having me. I love talking about building strong opportunities for science to make real world impacts. So let's get into it. I'm excited to tell you all my secrets.

Jason Stockman 1:16

Perfect. That's exciting. So, you know, maybe before we get started, um, could you share a little bit more about yourself for maybe people that aren't familiar?

Dr. Stevie Chancellor 1:24

Yeah, so I'm an assistant professor in computer science. That means I'm pre-tenure. Um, and I do research on AI and mental health. So I've been working in this space for over a decade at this point, trying to build more human-centered or people-centered solutions to problems that come up when AI intersects people going online to find content or support about mental illness. This is in two areas, social media and more recently, generative AI.

Jason Stockman 1:56

Well, let's dive in then. Um, it was great to hear your background. And, you know, with AI being such a hot topic, I honestly can't remember a single conversation with industry later where lately where they haven't wanted to dive more into that. So, you know, with everything going on kind of on the government landscape, you know, we've noticed that there's a lot more researchers that are interested in working with industry. But um, a lot of times they don't know how to begin. Um, could you share a little bit about your experience and kind of, you know, maybe your first major partnership, kind of how that actually began? You know, was there a specific gap in funding, just maybe more of a sudden interest in real world application? Or did you have companies just reaching out to you? How did how did it work for you?

Speaker 2 2:39

Yeah, I love talking about this because I think it's really refreshing to kind of pull the cape off of the box of how people get in with these like industry partnerships. I think people think it's kind of magic, but for me, it's been just over the years building up relationships with people that I end up closely just getting to know for good professional reasons. And um, there's an opportunity that makes itself really clear we should work together. So um in my PhD program, I'm actually interned twice at companies. Um, it's pretty common for computer science PhD students to take a research internship at a company. Um, you can think of the big kind of tech company, Silicon Valley, but also smaller uh startups will often have PhD uh research interns that work on product directly or that are doing research independently. And so I had two of those internships and was aware of how industry research operates and some of the incentives that they had through both of my roles. One of those was at Yahoo and the other one was at Microsoft. Um and then from there, the other thing that I that really kind of kick-started my connection are these relationships. So um I during my grad school met a lot of people at conferences and at events with the idea of just like getting to know other PhD students. And I wanted to just like know what people were working on because I'm a very extroverted person and people's stories really excite me. And my first industry partnership that really started to take off was actually a collaboration that we had with Cisco Research. And I used some of the skills that I had built for pitching at Microsoft for my projects at my Cisco pitch. Um, when they invited us, we had submitted a one-pager uh pitching our idea, and they invited us for like a 15-minute pitch, very short presentation to pitch a research idea for you know several hundred thousand dollars. And I used a lot of those skills for that first connection. That is not my biggest kind of collaboration to date, but it is definitely where I got started.

Jason Stockman 4:46

Okay. No, that's interesting to hear. I mean, uh, maybe this isn't a great question given kind of your history with industry and experience, but you know, as an academic, did you have any hesitations at all with this? Because, you know, we sometimes hear that researchers think they're maybe, you know, selling out um due to a lack of intellectual freedom. You know, did you specifically, or or maybe even address um if you didn't, you know, maybe why those concerns aren't maybe as valid as people think?

Dr. Stevie Chancellor 5:16

Yeah, so one thing that's important to me is the idea that my opinions aren't for sale. And so to me, there's a difference in selling out as in like compromising my integrity versus changing my opinions or perspectives on the landscape. And I really thought about this before I accepted a part-time role that I have right now at Google Research. So I am currently a visiting faculty researcher and I am part-time there. And I really thought about this idea of the integrity of my opinion and research. Um, and the reason that I was able to not come to terms with this too strong of a word, but like understand the role that I could have in the environment is uh in the environment of Google is partially because of the kind of impact that I wanted my work to have. So right now, large language models and the science behind them is dominated by industry, right? The places where you can get the biggest access to the best models and the frontier research is in companies like OpenAI, Google, and Anthropic. And I care passionately about making mental health and AI systems for mental health better. And I also care about protecting people's safety. And the team that I've joined at Google has been historically aligned with that perspective since its founding many years ago. It's been focused on the future of AI for a really long time. And the people on the team that I knew from these connections was really instrumental in talking to me about how to make sure that my work had an impact that felt authentic to my principles and morals. I'll be clear, I would be very allergic right now to working in industry full-time on AI modeling. I like my academic position and the freedom that it gives me. Um and I think that having the opportunity to make impact directly on products is a really great extension of my work.

Jason Stockman 7:21

So that leads really nicely into my next question when you talk about working in industry versus the freedom of academia. So, you know, when you think back to kind of the beginning, did you have any kind of culture shock in those first meetings? You know, obviously, as academics, we love freedom and deep nuances where, you know, industry is kind of looking for an answer yesterday. Um did you have to kind of navigate kind of we'll call it the different languages of industry and academia?

Dr. Stevie Chancellor 7:48

Not as much as I expected. So I anticipated having a set of product guidelines that I would have to hit, you know, really quickly and have pretty rapid deliverables. Um and I've been surprised at how focused this group is on getting the research done correctly. Um and I know a lot of times people feel like they need to trade the merits or the rigor of what they're working on versus the speed and efficacy that you need to operate on. And because the topic that I work on is so socially relevant and important that we get it right, which is dealing with um high risk or crisis situations when it comes to generative AI, um, everybody has been like, no, get it right. We we need to get this right rather than getting it done fast because they really want to make sure that the products that they're creating aren't hurting anyone. Um I will say I had to change some of my reporting strategies. You know, the idea of a loosey-goosey 45-minute update meeting that I would have with my uh PhD advisees, I don't have that with my manager. I'm very direct and get her information as quickly as I can. Um, and I've had to change some of my communication style. But for me, because the team I'm on is actually a pure research team, it's not a product-focused team, I've been able to have a lot more of the kind of culture that uh academia naturally um the culture that academia like naturally curates.

Jason Stockman 9:20

Perfect. Yeah, I know that makes a ton of sense. So, you know, it sounds like you have had a lot of success, which is great. I'm sure there's different secrets to that. I I maybe shifting a little bit away from your experiences, obviously building on them. But if you were talking to kind of more of a junior faculty or someone who's less experienced, um what would you recommend them doing to kind of vet out a potential industry partner? Um, you know, how have you kind of distinguished between maybe a partner that just wants an expensive, really smart consultant versus one that wants a true collaborator? Um, you know, how might you help the person that's new to this?

Dr. Stevie Chancellor 10:00

I love this question because I think it you're asking about the varieties of partnerships that you can use to build. And I think people assume that the way that you get an industry partnership is skipping from like zero to 60 miles an hour. But every good partnership starts with the underpinnings of the car that's gonna take you on the highway, right? Right. The reason that I was able to build up industry partnerships and collaborations is because I've built a network of people that are now themselves in industry and making decisions about who they want to work with.

Jason Stockman 10:33

Okay.

Dr. Stevie Chancellor 10:33

And I hate the idea of like networking for the sake of like networking, but like build relationships with people that you like, that have complementary views and insights to you. Um, identify people who are sharing the same kind of not only the problems they want to solve, but the worldview that they want to take to solving that problem, whether that's building a chat bot or a device. It's important that y'all, that you as a team are good collaborators, and the way that you build that is through trust and the relationships to start. Then I like to think about lighter, uh, lightweight ways to start to work with people. So I do have other collaborations that look more like consultants or um collaborate, lightweight collaborations on projects. Um, and we often start those off pretty lightweight. We share documents. First, we get an NDA in place, of course. Don't forget, don't forget about all the legal stuff. We got to get that in to protect all these like secrets. Um, but I like to start off lightweight. Can I give feedback on a white paper? Think about a project that you're about to start evaluating. Um, and then that lets you actually test to see if your relationship is good. So within the Google partnership, I actually was able to get a small gift from them through an open call that they had for funding, which Google has historically run these for a really long time. And I started working with two of their team members on the project that we ended up publishing as a paper. And it was great because I started working with people, learned about the approaches that they brought, was able to integrate a lot of their product insights that they had gotten from their experiences into the way we were doing our research. And it was like, oh my gosh, this collaboration is great. Let's take this to the next level. And that's when I got invited several years after that to um have this part-time role. The other thing, too, is these are not fast relationships, right? I don't think that you can expect to plant your like industry seeds and have them sprout in three months in most cases.

Jason Stockman 12:30

Sure.

Dr. Stevie Chancellor 12:31

These can take time to mature, and I would not rush finding a partnership for finding the right partnership for you as well as for your work.

Jason Stockman 12:43

Okay. Perfect. You know, um, you know, a lot of times with industry, they're they're looking to commercialize something, and and you obviously have interest in that, right? You want to see your technologies and ideas uh help people in the world. That's that's great. But you also have you know your scientific interests and responsibilities. Yeah. Have you had to balance those? And has that been challenging at all, or is it really kind of to your point about the relationships and the understanding of what you're doing together? How did how does how's that worked for you?

Dr. Stevie Chancellor 13:17

It really depends on the project that I'm working on and what the needs of the company are at a given time. So some projects are much more deliverable or product critique focused, right? Here is X product. Please go in and evaluate it for these Y things. I have a couple of those sitting on my desk right now that I need to do, and those are going to be a little more transactional. I'm okay with more transactional products and projects if A, I'm being compensated fairly for my time. I don't mind doing transactional things as long as I feel like I'm getting my time or my students' time well paid for. Um, but for the larger ones or the longer term ones, the important thing to me has always just been about alignment, right?

Jason Stockman 14:05

Sure.

Dr. Stevie Chancellor 14:05

I want to make sure that my ethos that I want to bring to my job works with the people that I want to work with as well as their demands. Like I'm not trying to get them to twist their arms to like not launch something or build a product if it's on their roadmap, if I'm part of that, but there needs to be good reasons for us to work together. And I'm tech is in this interesting ecosystem where interest in basic scientific research waxes and wanes, but it does have almost always some undercurrent, not undercurrent, it almost always has some interest in basic scientific research across the large tech companies. So a lot of computer science PhDs will go and work for tech companies to do basic science.

Jason Stockman 14:49

Okay.

Dr. Stevie Chancellor 14:50

Not basic, also applied science too.

Jason Stockman 14:52

Sure. Um you kind of read my mind there. My next question was really around the students, you know, how it sounds like it's it's definitely been a positive impact for you, but when you think about industry partnerships, what do you think it means for like in your case, your lab students, you know, PhD candidates, etc.? Um have you seen just some very high value in those areas?

Dr. Stevie Chancellor 15:16

You know, computer science maybe like other disciplines, I don't know about the rates of other PhD producing fields, um, but the majority of all PhDs produced in computer science in the US will go into industry. They will not become academics.

Jason Stockman 15:32

Okay.

Dr. Stevie Chancellor 15:32

And so it is to my benefit to be aware of what's going on in industry, but also to be able to connect my students with people who can help develop their careers. And so for my students, PhDs, my PhD students enjoy taking industry internships in particular as a way to meet new people, work on more applied problems, but also um develop the skills that they're going to need to translate the scientific skills to product innovations. Um and I'm very comfortable with them doing internships because they get to build that. And honestly, it's better for them to have a bigger and wider network. Now, the challenge I think comes when they are uh having to juggle their responsibilities with me as my RA or on their fellowships for their PhD progress versus their internships. And so we try to keep those very focused so that they can either focus entirely on their schoolwork and their science that they're doing at the U, or that they can entirely focus on their internship, like when they're off on summer internships are the big ones for us. We also sometimes will just bring industry collaborators onto projects if it makes sense. Um, and so working with people when they have the authorization to just like intellectually jam with us is a lot of fun. And you just like it's a really new, fresh set of perspectives that I wouldn't have had myself. And so the intern, the sorry, the industry collaborations I've had on papers where my PhD students have involved have been really fun for us. Again, you I keep underscoring this idea of just like enjoying myself. I like to think about most the best kind of work is going to be the ones that you enjoy the most. And waiting for those opportunities is, I think, the the strategy. That strategy that I prefer to take. I'm grateful I get a lot of opportun, not a lot of opportunities.

Jason Stockman 17:23

You're really segueing quite well into this. Uh, you know, obviously great advice on all the benefits and the you know the great experiences that you had in this, but I'm sure all partnerships aren't perfect, right? Um you know, have you had a a moment where maybe the partnership kind of hit a snag? Maybe there's a disagreement or a pivot. Um even if not, um maybe through your experiences or thoughts, you know, how how best should faculty handle those situations?

Dr. Stevie Chancellor 18:00

Yeah, I really think it depends on what the snag is caused by. So some snags are inevitable, right? With product direction changes or leadership change, those can be challenging to navigate because the control is out of your hands. And as a PI, I'm really used to having my hands on all of the science that comes out of my lab and being able to quickly adapt to these changes. But in these contexts, like I just can't, I can't stop my skip manager from changing directions one quarter. And that can be just really frustrating. And so I like to think about in what opportunities I have the capacity to change things myself versus things that are outside of my control that might be at the org level or at some level that's outside.

Jason Stockman 18:49

Sure.

Dr. Stevie Chancellor 18:50

And then I think about so let's assume that there's some things you can't control, and those you just have to kind of tolerate and deal with, which is uncomfortable, but kind of how things naturally go. But then I think about the ones that I have disagreements over, and oftentimes it's because of not the speed of what we're working on, but what the outcome is going to be. So even though I'm lucky to work often with researchers who don't necessarily have to deliver on product, they're incentivized to improve product at some point or to be able to socialize other product groups to do more safe or impactful work with AI mental health. And sometimes our disagreement is not about what we find valuable, but what we have to prioritize in a given moment to proverbially scratch each other's backs. So I think a lot about developing maybe some opportunities where I may give a little bit more that's less interesting for me. But my hope is that in the future they're gonna come around circle around and be interested in the things that are, you know, valuable for me. And I admit that like part of the the best part about working in industry, especially about AI right now, is that the work I'm doing can have direct impact on products. So I could just say, I don't want to write new papers, I don't want new patents, just let me put my hands on the model and fix this. And I can allow to do that as in my role. And that's something that I'm looking forward to doing in my position that I've only been in this position for like six months. So we're still seeing how it's evolving and changing. Um, and my manager is really supportive if I was just like, hey, I want to, I really was want to stick my hands in the model. Um, and so in terms of resolving it, I just think, okay, I keep coming back to relationships and how important those are for building up um a structure for an industry academic partnership. Like the partnership is really important. And I think people kind of skip over that part when they're trying to find people or companies to work with, right? They pull like a huge list of oh, the top X companies in the area or in their you know research space, and they're like, oh, how can I get in at you know, this company or that? But like part of getting the impact that you want with your science is remembering that the science has to be in a place where it will thrive, and you can't force that at every opportunity on that list.

Jason Stockman 21:16

Sure.

Dr. Stevie Chancellor 21:16

So for me, it's finding the right people at the right time that are going to be down to do to take my ideas and innovations and pull it uh forward.

Jason Stockman 21:26

Okay, great. Um, you know, something we get sometimes this is kind of my last like long-winded question, then maybe we'll get into a lightning round. But um sometimes we get early career faculty um that are worried that their industry work wouldn't count towards like tenure or promotion. Um, what advice or what would you tell those folks?

Dr. Stevie Chancellor 21:47

I do not have tenure, nor am I promoted yet. So gigantic grain of salt. Um, you know, I think that first of all, get advice. From senior people in your area. That not only includes the people in your department, but also people who might be writing letters or supporting your reappointment or tenure and promotion packages. They will have a much better sense about the amount of time, effort, and outcomes that are important for you to prioritize in building up a case so you can be successful at tenure. In some fields, that may look like patents or discoveries. In other fields, it may look like, hey, you shouldn't spend more than X, you know, amount of time in a given year at a company because of the risk that it has to your tenure package.

Jason Stockman 22:37

Okay.

Dr. Stevie Chancellor 22:38

The other thing that I think is is valuable to consider what's the right way to put this? Sometimes it is okay if you have to wait to really act on this until there is a time that's good for you in your career. And so there are some benefits to chasing heavier partnerships once you have tenure or you have promotion or more job security because you can take a little bit more risk. You have a little bit more freedom with leaves of absences or alternative appointment structures, going on sabbatical, being in a place physically is one really awesome avenue that I know people use to get into industry for a break, but also as like a really refreshing um way to think about science from a different impact. So my advice is basically ask and make sure that what you're doing will be safe and good for your tenure. And if it isn't for you right now, it's okay. Find other avenues to bring this into your work at a future date.

Jason Stockman 23:43

All right. Um you up for a quick lightning round, and then we'll finish up.

Dr. Stevie Chancellor 23:48

I am.

Jason Stockman 23:49

I think you kind of answered this a little bit, but just quickly, you know, what do you think the best way is to find or meet an industry partner?

Dr. Stevie Chancellor 23:57

Get in front of them. Get in front of them and events that they're at where they're gonna want to talk to you. Conferences, industry, tech conferences, events at the U, get in front of them.

Jason Stockman 24:07

Great. Um, this might be a hard one, but uh, you know, think back 10 years ago. What is one thing you wish you knew about partnering with industry? Maybe it was five years ago, six years ago, whatever. But what is one thing you wish you knew before you started out on this adventure?

Dr. Stevie Chancellor 24:24

Be straightforward about what value you bring. Don't tangle up the in value of your scientific impact in long academic language. Speak simply, speak clearly, and speak to impact.

Jason Stockman 24:40

You might have just answered my next one, but uh Oh, great. Do you uh what do you think is probably the most common mistake faculty make when pitching companies?

Dr. Stevie Chancellor 24:51

I think faculty should reverse their pitches. So lead with what you do and why it matters and how you're gonna do it rather than the standard academic model of what it is and how you did it, and the why always ends up at the end. Don't do it like that. I've made that mistake more times than I can count.

Jason Stockman 25:10

I love that answer. That's a great one. Uh, last question. Um do you think there's any ways that uh us here in the corporate engagement center can best support faculty that are interested in pursuing industry partnership opportunities? You've probably had some experience with us, but maybe not a ton. Is there anything that comes to mind?

Dr. Stevie Chancellor 25:30

Two that are big. One is smoothing over legal and sort of contractual barriers, getting us in front of people who understand the industry that we're working with and why they're asking for certain structures of NDAs or contractual agreements. Uh, dealing with legal quagmiers is one of my least favorite things about working with industry.

Jason Stockman 25:51

Sure.

Dr. Stevie Chancellor 25:52

Um, and the headache of having to share information between the me being in my U role versus me being in like my, you know, my industry role.

Jason Stockman 25:59

Yep.

Dr. Stevie Chancellor 25:59

That's one. And then the other one is helping to identify partnerships that matter for more than just me. So I can list off a bunch of companies that might be interested in relevance to my research.

Jason Stockman 26:10

Sure.

Dr. Stevie Chancellor 26:10

But I think that the you can capitalize on knowing lots of faculty that cross-cut areas. Say, for example, AI and society. Um, I don't know anything about polymers, but I assume there's a bunch of people doing cool stuff about polymers, right? There are, yes. You can have that bird's eye view that I can't have.

Jason Stockman 26:30

Gotcha.

Dr. Stevie Chancellor 26:30

And making those connections to find the right companies is is valuable for me.

Jason Stockman 26:34

Awesome. So, Dr. Chancellor, um, this discussion's been incredibly lightning, and I love how positive it was. So I really appreciate that. I I do want to thank you for sharing both your thoughts and your experiences with our fellow researchers here at the university. And really, our goal is always to help share insights and experiences to help anyone interested in industry partnerships feel better equipped to go out there and and obtain their own. And I do feel like we we definitely accomplished that goal today. So, again, thank you so much for joining us.

Dr. Stevie Chancellor 27:04

Thanks for having me. I love chatting about industry folks or industry partnership. And anyone wants to chat more, I'm happy to happy to chat.

Jason Stockman 27:11

Sure, I'm sure there will be plenty of people take you up on that. So, so with that, um, thanks everyone for joining us today on Demystifying University and Industry Research Partnerships. You're listening to Bridging the Gap, where we help build a bridge one partnership at a time.