In this episode, Dimitrije Jankovic shares about his role at Sanofi and how the organization is working to figure out a way to use generative AI responsibly, as well as Sanofi’s plan to pave a way to data democratization not only for Sanofi, but for all of their end users as well.
In this episode, Dimitrije Jankovic shares about his role at Sanofi and how the organization is working to figure out a way to use generative AI responsibly, as well as Sanofi’s plan to pave a way to data democratization not only for Sanofi, but for all of their end users as well.
---
How you approach data will define what’s possible for your organization. Data engineers, data scientists, application developers, and a host of other data professionals who depend on the Snowflake Data Cloud continue to thrive thanks to a decade of technology breakthroughs. But that journey is only the beginning.
Catch up on the latest announcements from Snowflake Summit, including advancements with generative AI and LLMs, flexible programmability, application development, and much more.
Watch now here.
[00:00:00] Steve Hamm: Dimitrije, it's so great to have you on the podcast.
[00:00:03] Dimitrije Jankovic: Thanks for having me.
[00:00:04] Steve Hamm: me. Let's get started here. Uh, I, I think it would be great if you'd start by describing Sanofi's organization and its business kind of broadly, and tell us what distinguishes the company from other large pharma companies.
[00:00:20] Dimitrije Jankovic: Absolutely. So Sanofi is the world's largest vaccine manufacturer operating at over 90 countries and providing healthcare solutions to over 160. Now we do everything from molecule through to market. Meaning that we operate in vaccines, specialty care, general medicines, and consumer healthcare. We have over 60 manufacturing sites in 30 countries and over 20 R&D sites, like I mentioned, molecule to market.
It means that we do research and development, manufacturing and supply chain, commercial and medical. We really do it all, all over the world and we're a large organization, but what's gonna help to continue propel our growth. Is actually our innovation and the ability to tap into the power of data and AI and connect the dots in a way that's never been done before.
[00:01:06] Steve Hamm: Yeah, so it sounds like a really important part of innovation. There is not just. With molecules and, and processes and stuff like that, but also with, with technology, IT, and, and data. So that's really interesting. Now, I understand that the company is kind of three years or so into a major digital transformation.
So please tell us what gave rise to the initiative and describe the strategy and goals and, and maybe give us a, a bit of a progress report.
[00:01:35] Dimitrije Jankovic: Absolutely. I mean, fundamentally we were looking to move from seeing digital within the organization as an enabler to seeing digital as a strategic partner. So that meant that we're gonna come together as one and streamline our processes, provide scalable platforms, and allow teams to build differentiating products that transform the way that they approach a problem within their business function.
Now fundamentally, our, our digital transformation is hinging around being the leading digital healthcare platform. Now, what that means is that we want to be able to provide the backbone, the foundations, and the products that allow us to transform how we interact with all of our stakeholders, whether they're scientists, patients, folks within the manufacturing sites.
We're really looking to affect everyone Now. Our strategy remains ambitious, so there's still a way to go. We've done a massive amount of work in terms of leapfrogging ahead of where we were couple years back, but this key to success is gonna be to continue investing in our people and being clear about how we not only enable our specialized services, but also transform the lives of our everyday Sanofians.
We wanna enable the masses through digital, and that means that we give them access to the power of data and ai.
[00:02:47] Steve Hamm: And by the masses you mean the masses within your workforce and partners, but also the masses within kind of at the customer end.
[00:02:55] Dimitrije Jankovic: A absolutely. I mean, if you think about an organization of our size, we've got over a hundred thousand employees, and that means that not only 2000 or 3000 or 10,000 people within the organization should be touching, uh, data each and every day, and making decisions using advanced analytics and ai, that it really means that everyone within the organization, Has access to this and it's pivotal to their day-to-day success.
[00:03:20] Steve Hamm: Yeah, yeah. Now your title is the Head of Data and AI at Sanofi. What role are you playing in the transformation?
[00:03:28] Dimitrije Jankovic: So my team and I ensure that we're forward looking in the sense that we're defining our positions or stances on important topics such as responsible AI and where we would want to implement generative ai. Think of that as sort of setting the North Star to some extent. Now, once you've set the North Star, you wanna figure out within the next 18 months, what are we actually going to do as an organization?
And probably just as importantly, what are we not gonna do to move towards that North Star? So that includes the delivery of key strategic initiatives and really driving forward the transformation from a data and AI perspective, meaning that we can work on re-skilling our organization and increasing our data quotient all the way across.
And last but not least, I mean, we've actually gotta deliver on that strategy. So the point is that we're actually selecting the programs that we're delivering against being able to course correct and make trade-off decisions throughout the year.
[00:04:24] Steve Hamm: Yeah. Yeah. Now you used the term responsible ai. Um, I know I, I've, I've heard it. I have a sense of what it is. I'm not sure all of our listeners are familiar with it. So would you define that and say how you're applying it there at Sanofi?
[00:04:41] Dimitrije Jankovic: Absolutely. So the topic of generative ai, I think has brought forward the topic of responsible AI across a number of organizations. Now, many organizations have done something about this in recent years, but if we really think about it, it comes down to what do, what safeguards and controls do we need to have in place as an organization to ethically and responsibly implement AI at Sanofi?
That means that we're, we've named it RAISE. Which means that we're looking at responsible AI at Sanofi for everyone. And I wanna highlight that element of for everyone, because everyone is gonna be making decisions using systems that are powered by ai. Everyone is going to be accessing and have access to these tools.
And so fundamentally the way that we look at it at Sanofi is that this is not a digital only problem. It's one where we need to closely collaborate with privacy, with procurement, with cybersecurity. With, uh, legal, uh, with our IP team, there's a variety of stakeholders here who are key in helping us set forward our strategy around responsible AI and then implementing it, finding the right training controls, uh, governance and technology to power it.
[00:05:53] Steve Hamm: So do you have some kind of like cross corporate committee that's looking at responsible ai or is it just kind of something that everybody's supposed to be thinking about?
[00:06:03] Dimitrije Jankovic: Well, we've set forward a working group that's helping us define the strategy and, and define exactly what responsible AI means at Sanofi. So we've got a, a set of components. It's influenced by our internal stances, some regulations and external leading practices. But to really bring that to life, we've got a responsible innovation governance committee that's gonna be making some key decisions associated.
It's about, and it's spearheaded by our chief digital officer as well as our, uh, chief, uh, legal business integrity officer.
[00:06:34] Steve Hamm: Very good. I, I'm hoping that like every company in the world is doing this because it seems like it's really gonna be necessary. Hey, you know, in, in, in regard to ai, I, I've, I've heard the term, you know, that you guys are building an AI factory. Is, is that a good term to describe what you're doing?
And, and in that evokes kind of certain images? What, what, what do you mean by that?
[00:06:58] Dimitrije Jankovic: Well, I, I guess the way that you explain AI changes month in and month out. So a couple months back, we thought a good way to explain what we're trying to do is to focus on a factory more so from the elements that it's repeatable and scalable, right? So if you think about ai, AI has been within organizations for a number of years, but often where it's been happening is in.
Sort of pilots in certain applications where you're making a one-off decision. And what we want to be able to do is either develop AI products that are completely unique and kind of allow us to do the ever been done before or they affect scale, right? So I mentioned that we operate in a ton of different markets.
You know, you want to be able to build a tool out that gets implemented across the globe. And so the concept behind a factory was really that we enable scale with consistency.
[00:07:46] Steve Hamm: Yeah. Very cool. I I get that. Um, hey, you know, I, I checked you out on LinkedIn. I do this with all of our guests and it strikes me that you are quite young. To have such an important role in a giant global corporation, how did you get so far, so quickly, and what lessons have you learned along the way that are helping you deal with the challenges in your current job?
[00:08:11] Dimitrije Jankovic: I, I selected sort of focusing in on data and AI fairly early in my career, and I, to some extent, I've had the benefit of data and AI really lifting off over the past couple of years. But above all else, I've had incredible luck in my career to work with amazing leaders who've given me space, taught me some important lessons, and provided incredible coaching.
So, I mean, a few important lessons that I've learned over the years is to always start with why before jumping into a solution, and I'm an engineer by training, so I, I tend to do that, is to really think through the problem statement itself, ask for advice. I mean, to this day, I have regular retrospectives and coaching with not only my leaders, but also my colleagues and peers so that they can share their experiences and how I can learn from their experiences and how I can improve myself. And, and last but not least, I think, is to not be afraid to get your hands dirty and and fail. My teams tend to operate on a basis of radical transparency. Um, and that's extremely important to build a sort of a nurturing environment where everyone can grow.
[00:09:11] Steve Hamm: Yeah. Yeah, so people feel comfortable making mistakes and failing as long as they learn from them and, and correct them, and are open about them. So I think that's a very healthy culture. So congratulations on that.
I know that this started a bit before you joined Sanofi, but please explain when and why the company began to move to the to Snowflake's data cloud.
[00:09:37] Dimitrije Jankovic: Well, fundamentally, as an organization, we want to put the power of data into people's hands, and that means to us that we're on a path of data democratization. So we believe that every employee at Sanofi can use. Data and should have different ways of engaging with data. Now, we needed a simple way to get access to scalable compute and storage without a ton of overhead so that not all our users need to be developers to get the value that they need from data.
And that's ultimately where Snowflake came in. I mean, it allows for high self-serve capabilities users to be enabled without a ton of administrative overhead, and it allows for computer that's scalable and storage that's scalable. So, The backend platforms really from our perspective, need to be abstracted to our end users.
They don't need to know what's happening behind the scenes, similar to a factory, uh, but they need to be able to achieve the end outcome that they need.
[00:10:30] Steve Hamm: Yeah. Yeah. Now I, I've heard the term data mesh in terms of what you're doing, and I, my sense is that different people mean different things. When they say data mesh, what do you mean by it there, and how does, how does the Snowflake technology fit in?
[00:10:44] Dimitrije Jankovic: Yeah, well at at Sanofi, we're looking to drive that democratization by giving access to the right platforms, to the right people at the right time, with the right processes, right? And so Snowflake allows us to create what we're calling our data foundations across business domains. To create our data mesh.
So it allows for our data assets really to be discoverable, addressable, self-describing, trustworthy, secure, and interoperable. These are core tenants of our data mesh itself, and Snowflake is a key portion of that architecture.
[00:11:16] Steve Hamm: Yeah. And the data mesh, it, it sounds like is, is something that allows people, it kind of cuts across business units or, or functional units, right. So that, so that data isn't stuck in silos.
[00:11:30] Dimitrije Jankovic: E. Exactly. I mean, the data mesh, the way that we're looking at it is that technology is only a portion of it, it's tech enabled, but fundamentally, this mesh allows for data to be reused, maintained, and governed. By all Sanofians, whether they're data scientists, analysts, engineers, or end users within a lab, each person needs to be enabled by the mesh, and each person will have certain responsibilities in terms of how they manage their data within the mesh. So it's sort of a give and take model.
[00:11:58] Steve Hamm: I understand that you're using Snow Park, which is one of, of snowflake's core technologies in the data mesh. So explain, you know, please explain. You know, what it is, how you're using it, how that's going.
[00:12:11] Dimitrije Jankovic: So Snow Park is a platform or a capability that Snowflake has enabled, which allows our developers to. Build their data pipelines and build their machine learning algorithms in the languages of their choice. Some of our folks like to use Python, some of our folks like to use sql, so it really allows for more flexibility in terms of how we use Snowflake.
Originally, we did it as a mechanism to simplify our architecture, reduce some of our operational costs and administrative burden, and, and really fundamentally tap into the scalability that Snowflake off offers to improve end user, um, experience. So, And now we're looking at some more advanced use cases, moving beyond just sort of that wrangling and crunching of data to potentially using it for advanced analytics and ai.
[00:12:58] Steve Hamm: Okay. Very cool. Now, in your, in your user interface for your mesh, for, you know, all those a hundred thousand people who are tapping into it, you have a couple of interesting features. There's one search and one ai. Tell us about those.
[00:13:15] Dimitrije Jankovic: Absolutely. So, OneSearch, you can think of that as a concept that we came up with to make our data more searchable and accessible. So by virtue of being such a large Com large company, we didn't have the means to really centralize all of our data, nor do I think that that's necessarily the approach that we wanted to go after.
But we wanted to make sure that anyone at Sanofi can search for data, comment on data, request data, and have access to it through a common front end. So this product is. Kind of been released to the organization with the idea that it is a minimum viable product that we're gonna discover more and more data needs.
And the data that's the most needed ends up being the data that gets brought onto, uh, one search.
[00:13:54] Steve Hamm: is one search, a very simple interface. Like a, like a Google search
[00:13:59] Dimitrije Jankovic: Yeah, it's, that's actually the way that we sort of, Propose it to the organization. It's, it's similar to Google in the sense that you've got first things first, just a search functionality. And now we're looking at saying, okay, well, similar to the way that Google did Google images, over time, you should be able to actually search for dashboards, right?
Or you should be able to search for a clinical study so that there's a variety of capabilities that we want to add to one search, but it's a very clean sort of front end. Very similar to Google. I'd say we're, we're pretty inspired by Google.
[00:14:31] Steve Hamm: Oh, yeah, that's easy to be inspired. So you have a couple of different filters basically for, for the, for the general user. And that sounds great. Now go on, tell us about one ai.
[00:14:44] Dimitrije Jankovic: So while if you think of one search as everyone needs to get access to data one, AI is really the specialized platform that we have for. What I'll call industrial scale ai, and the concept here is that our data scientists need to have an environment or a workbench in which they can develop code and then actually publish it out into a factory where those use cases can be repeatable.
So it's, you know, our philosophy is that not everyone needs access to a one AI sort of tool set, because not everyone is that kind of user with the data. Some folks will just need access to an environment where they'll be able to discover. Whereas others may need access to what we'll call, uh, a more high performance area for research and development, manufacturing and supply chain, and developing some of those AI products within our factory.
[00:15:36] Steve Hamm: interesting. So one AI, and correct me if I'm wrong, is kind of a catalog that lists the products created by the factory.
[00:15:44] Dimitrije Jankovic: Yeah, the one AI is the environment where developers build these products, and then the concept is that once you've developed it, you can actually reuse it in different areas, so you can publish that, people can reuse it, and you can also use it and scale it out to different markets, sites, et cetera.
[00:16:01] Steve Hamm: Got it. That sounds really cool. Now you, you said kind of upfront when we started talking about the data mesh, that it's kind of a work in progress, but it'd be great if you could give us a status report, uh, on the initiative, you know, what benefits is the company getting from it, and kind of tell us where you're going next.
What are the next steps?
[00:16:22] Dimitrije Jankovic: A Absolutely. I think one of the main benefits has been associated to the mindset shift. So once the strategy's been set and the architecture has initially been developed, we identified some gaps where, you know, different areas of the business are using different tools. We identified some gaps in tooling that we have altogether.
But now it's really about meeting supply and demand, and where we can prioritize the actual technical features we need to build out in the mesh based on what the end users are actually leveraging. The biggest portion that I think was neat about it was the complete mindset shift. That data is no longer just tied to one source system, and that's the only way that you use it, but rather that data is tied to a business domain and you can reuse that same data asset in different areas.
That really comes down to just common processes and, and governance, which is to some extent some of the, uh, the hard work that goes on behind the scenes to make the mesh a reality and, and having really formalized roles
[00:17:19] Steve Hamm: That's really interesting because, you know, it, it's a great idea, it's a great strategy, but it just doesn't, you don't just, you know, snap your fingers and make it happen. There's a lot of kind of blocking and tackling to, uh, to get it done so,
[00:17:33] Dimitrije Jankovic: Exactly.
[00:17:34] Steve Hamm: Uh, well, when you look ahead, I mean, if you look at this whole initiative started three years ago, um, as part of the, the transformation, what, you know, I mean, is this something that just continues into the, into the future or is there kind of like an end point?
Oh yeah, we that, we got that done. That's what we, we intended, then we turned to something else.
[00:17:54] Dimitrije Jankovic: Well, the beauty of the mesh is that the end users actually drive the prioritization of what happens next, right? So the concept behind it is that you prioritize the maintenance of the data and the governance efforts on the data that is most used. And so if you end up needing new data for new use cases, you're tackling, particularly with the emergence of.
Uh, generative ai, I mean, there's more and more appetite for it within every single organization. Well, that'll allow you to identify what data do I actually need to execute on that. I mean, there's the common saying, and I'm sure we've all heard it, of garbage and garbage out, but it really does come down to how do we actually set up the common processes around making sure that that data is not only cleaned once, but it's actually sustained on an ongoing basis, um, to be usable by the organization.
[00:18:43] Steve Hamm: Yeah. Yeah. Very cool. Now, you've mentioned generative AI a couple of times. Just to give some context, you know, over the past few years, all sorts of organizations have been using machine learning to improve their data models, to empower employees with data. You know, very, it, it's been a powerful force, but all of a sudden, These large language models, the generative AI applications, these are powerful new tools, and they're becoming available to organizations and they're being, you know, they're, a lot of people are, are, are making 'em available to a lot of their employees and, and, and even customers and things like that.
So can you name like one or two specific application that you've, that you've already got in place at Sanofi?
[00:19:27] Dimitrije Jankovic: A absolutely. I mean, large language models have already been utilized at Sanofi prior to this, within the space of, uh, drug discovery because there's a lot that can be done with large language models There. Uh, but now it's really about, you know, as more and more partners are coming to market with solutions, being able to cut through the noise and identify really the use cases of largest value.
So we know that there's some common capabilities that we're gonna want to have as an organization in terms of content tagging, tech summarization, knowledge extraction, but then it's really applying those within research and development or manufacturing and supply chain or commercial and real world data.
So, Each of them will have their own unique needs, and so we're gonna have to identify the real use cases that we want to double down on.
[00:20:11] Steve Hamm: Yeah. Well, I'm gonna press you a little bit more. Everybody wants to hear about some really cool gee whiz technology. Some application that's already at work or that, or that you're, that you're, you're gonna bring to market internally very soon. Can you give us one of those little, a little for foresight?
[00:20:28] Dimitrije Jankovic: Absolutely. I mean, one of the things that we've been looking at is, Really innovating in the space of R&D. Now, I'm not an expert in R&D, I'm not an expert in computational sciences, but I, I have colleagues that really are experts in this space. And some of the things that they've been able to do is generate active learning models that they've taken from mRNA, applied it to large molecule research, applied it to small molecule research, and it's, it's actually a product that we've built out internally, uh, called Alien. So active learning,
active learning fundamentally looks at how we can change the way that we tackle the training data that we provide to our machine learning models and inject it with. Uh, actual outputs and create a feedback loop between what the model has generated and what our training sets do, so it really makes the model smarter over time.
[00:21:22] Steve Hamm: So the model kind of becomes a better and better match with the reality that's out there.
[00:21:28] Dimitrije Jankovic: correct and it consistently has a human involved in the loop so that you can actually compare the way that you used to do things, along with the way that the model is, what predictions it's making, and it's particularly effective in target identification and, and drug discovery.
[00:21:44] Steve Hamm: Yeah. Yeah, I get that. That's very cool. Now we, we've talked a little bit about generative ai. You know, some of these advances that have come in the past. Hell, it's only been six months. Really have been. It's fast and furious. Seems like every time you turn around, there's another article or, or for people like me, we, we, we consume this in articles.
You, you consume it in your, in your daily life, but it seems like. Innovation has just accelerated. I mean, this is just a crazy, crazy time. So, and, and it's even harder to tell how, you know, it's probably harder to predict what's coming next, but I'm gonna ask you to do so anyway. So looking out over the next year or so, what do you think will be the major emerging trends in data management analytics, AI?
[00:22:34] Dimitrije Jankovic: I'd say it's a good question. I think, uh, and I heard this at a conference not so long ago, but that the. Pace of change that we're experiencing today is the slowest that it'll ever be. Meaning that things are only gonna get faster ahead of us. And so I, I think over the next year, one of the key trends that's gonna exist is actually cutting through the buzz in generative ai because everyone, like you mentioned, there are articles that are all out there.
I mean, chap, GPT has final, finally made AI less intimidating and accessible to all. But that also means we need to focus on making sure that we're upskilling the organization, providing the appropriate training and change management, and designing a solution landscape where our generative AI can fit.
So I think the emerging trend for me isn't just about the neat applications, but it's how does it fit within the organizations themselves and those that figure out how to adopt it are really gonna have a competitive edge.
[00:23:28] Steve Hamm: yeah. Yeah. It's not gonna be a one size fits all. There's all gonna be a lot of, of tweaking and customizing and choosing it sounds like. Yeah.
[00:23:37] Dimitrije Jankovic: I think it's selecting the right partners, fitting in that element of a responsible AI strategy, and finding a good way to optimize your total cost ownership. Cuz sometimes you're gonna solve problems that might not be worth solving.
[00:23:50] Steve Hamm: Yeah. Well, let me see if I can stump you with the next question. So looking out over the next five years, you know, put on your visionary cap for a couple of minutes, how do you expect data technologies and AI to change business and even society?
[00:24:09] Dimitrije Jankovic: I think that fundamentally one of the elements that is gonna be huge in the data space is actually enabling better data sharing between organizations, whether those are two private sector organizations. Private public partnerships, ultimately creating a, a more cohesive data ecosystem where you can train models on data sets that are outside of your walls and that becomes the norm, and it will, at the end of the day, I think, put the hands of, or put the data in the hands of the patients and of the end users.
I, I think that five years from now, the way that we share data and the way that we approach data in between organizations and as individuals is gonna fundamentally shift.
[00:24:52] Steve Hamm: Put some color on that. How will it be different?
[00:24:55] Dimitrije Jankovic: I, I, I think it'll be different in the sense that it'll be a topic where there's a lot more controls that are in place from a technology standpoint, and there's a lot more regulation. And so you're gonna have to learn how to navigate the regulatory space. So making sure that you know what we're doing is ethical.
What we're doing is, um, compliant. While still enabling more and more partnerships. So if we think about health data, it's really about how rich that data set is. And a lot, a lot of that comes down to having access to historical data. And fundamentally, a lot of that is within academic institutions. It's within, uh, different sorts of organizations and figuring out how to share that data and getting more and more personalized data from patients moves us towards per personalized medicine. But you need to have the right infrastructure and the right processes and governance in place to do that.
[00:25:46] Steve Hamm: Yeah, absolutely. You know, it's interesting, you know, one of the great growth professions over the past, well, it's been, it's been 50 years, but it gets faster and faster and more and more is software programming, yet ai. You know, is already being used to do a lot of basic coding and it seems like it's gonna do even more in the future.
So when you look to the future, how do you see kind of within your own organization, your own tech organization, is there gonna be as many people, are there gonna be different kinds of jobs because you know, of like copilot and some of these other kinds of, of AI automation and programming?
[00:26:28] Dimitrije Jankovic: I mean, you asked me what I thought the major emerging trend in data and AI would be in the next year, and I answered that the focus would be on actually up skilling the organization, providing training and and change management, which probably isn't the most sexy of responses. But I do think that, like you said, you know, there's stats out there that are showing the programmers can be 60% more productive if they have access to a tool like copilot.
And so from my perspective, it's really making sure that the people that are feeding. These, um, generative AI models have a very, very strong basis of programming principles so they know what they're doing, uh, and know how to use these tools within their day-to-day lives. And similarly, it's gonna transform a number of professions, and it's about how we embrace it and create a more productive workforce.
[00:27:20] Steve Hamm: Right, right.
Now we typically end the podcast on a more personal note, kind of something a little offbeat. I see that you opened the office for Sanofi in Toronto, the AI Center of Excellence. So why Toronto? Is it just because you live there or, or is there, you know, another reason why that's specifically a great place to have this?
How has this, you know, how, how did it come to be and how is it developing?
[00:27:48] Dimitrije Jankovic: Yeah, I'd, I'd like to think that I was the driving force behind the selection of Toronto, but I, I, I fundamentally, it was not, um, I was employee number two in Toronto, so I did have the benefit of opening the office in Toronto. And if, if we think about Canada in general, Canada, between Toronto, Montreal, and Alberta has a lot of.
Very strong AI talent. And if we just look at and step back, what industries does AI have the most impact in? You'll often see across every industry report that it could be in healthcare. And so by virtue of having a very strong talent pool and having, you know, uh, uh, sort of niche that we can tap into, we focused in on getting talent that wanted to make a difference in the world, uh, using data and ai. And that's why we selected Toronto and, and we've grown at a pretty rapid pace here.
[00:28:37] Steve Hamm: How many people do you have now?
[00:28:39] Dimitrije Jankovic: W uh, so we opened the office officially, no, it's not a state secret. We o opened the office officially, uh, just over a year ago. So a year and I believe nine or 10 days. And we've got 107 staff today. We've got some amazing interns in those numbers. But, uh, the, the team is growing and we've got our first cohort of new graduates joining just next week.
[00:29:03] Steve Hamm: Oh, cool, cool. And interns don't just fetch coffee anymore. They're,
[00:29:08] Dimitrije Jankovic: No.
[00:29:08] Steve Hamm: they, they're really ex essential players,
[00:29:10] Dimitrije Jankovic: there, it's, it's incredible to see what, what can happen once you give, um, some folks some support and some training and access to the right sort of problems. It's, it's really, I, I think, one of the most motivating parts of my job.
[00:29:24] Steve Hamm: yeah, yeah. Hey, talk to us about the Catalyst program. What is it and what's it supposed to do?
[00:29:30] Dimitrije Jankovic: So I, I think just like many new grads, when I came outta school, I wasn't a hundred percent sure exactly what industry I wanted to focus in on or, or what, um, portion of data and AI I wanted to focus in on and my or earlier career allowed me to select and pick and choose and try different things. And so the Catalyst program is really allowing folks to join Sanofi.
And try different elements of Sanofi. So, you know, not everyone at a school knows that they want to do supply chain or that they want to do, uh, development, or that they wanna work in real world data. Some folks do, but others might not. And what it allows 'em to do is have a rotational program over the course of two years to try different rotations, try different programs, and then find what it is that really is their passion area.
And so we've got catalysts in data science and data and business analysis and data and AI strategy in, uh, data and software engineering. So we've really got a variety of roles and they get to actually try different things in different areas and, and find what it is that makes them tick.
[00:30:33] Steve Hamm: Yeah, that sounds like a very good way of meshing people, you know, their, their skills and capabilities with what you need and fits in with what you talked about before about how evolving the organization things are, changing fast new roles, and getting people in who are very kind of both flexible, but also start to get a sense of what they're really good at.
I, I think, is gonna be really good for the organization and also good for them.
[00:30:58] Dimitrije Jankovic: Absolutely. It provides them a, a good start to their career. And it also allows, I mean, a lot of the technologies that we're using, like you said, haven't been around for all that long, so it allows them to sort of really go in as a clean slate, learn a lot, do a lot, and, and have an impact on the world.
[00:31:16] Steve Hamm: Yeah. Yeah. Well this has been a really interesting conversation, really fun and kind of challenging. Uh, it seems to me that your company is almost like a case study for how big organizations can. Be on top of the, the, you know, the newest technologies that are coming out. Figuring out how to use them, how to use them, you know, not just kind of in a generic sort of way, but in a very prescribed sort of way.
And I really love what you told us about responsible AI and how you're really, the company's really thinking about this very deeply thinking about safeguards, controls to make sure that. This new technology is used ethnically, uh, ethically and responsibly. Uh, I love the fact that you've got a governance committee.
I think every company should have it. And do I think every company should do what you, you guys are doing? Cuz I think this is very powerful technology and it has to be used very carefully. So I, I congratulate you guys on that.
[00:32:18] Dimitrije Jankovic: Yeah. Th thank you. And I, I think, you know, at the, the pace at which this is evolving, my advice to anyone would be to really just get started with a minimum viable product. Like, it's not gonna be ideal. It's not gonna be perfect off the get go. And that's one of the, uh, transformations and thinking that our organization has taken in terms of getting started with responsible ai.
[00:32:38] Steve Hamm: Good advice from a young but wise man. So thank you very much, Dimitrije.
[00:32:43] Dimitrije Jankovic: Thank you Steve. Thank you for having me.