The Data Cloud Podcast

Change is the Only Constant with Omar Khawaja, Global Head BI Roche Diagnostics at Roche

Episode Summary

In this episode, Omar Khawaja, Global Head BI Roche Diagnostics at Roche, discusses the intersection of healthcare and big data, how to handle the balance of centralized and decentralized data, the future of data sharing, and much more.

Episode Notes

In this episode, Omar Khawaja, Global Head BI Roche Diagnostics at Roche, discusses the intersection of healthcare and big data, how to handle the balance of centralized and decentralized data, the future of data sharing, and much more.

--------

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.

Attend Snowflake Summit 2023 in Las Vegas June 26-29 to learn how to access, build, and monetize data, tools, models, and applications in ways that were previously unimaginable. Enable seamless alignment and collaboration across these crucial functions in the Data Cloud to transform nearly every aspect of your organization.
Learn more and register at www.snowflake.com/summit

Episode Transcription

[00:00:00] Steve Hamm: Well, it's good to have you here, Omar. Um, you know, Our listeners have heard of Roche, but they may not know that it's the largest pharmaceutical company in the world. and that the name Roche is short for F Hoffman Laroche, AG. And that it owns a much more familiar name and U.S. Biotech Genentech. So if you would please begin by describing the scope of the company and then focus in on Roche Diagnostics, where you work.

Omar Khawaja: Hi, Steve. Thanks for having me on the podcast today. Really appreciate it. Pleasure to be here. Uh you're right. Uh, Roche is a leader in pharmaceutical healthcare industry. Uh, We have two big divisions: pharmaceuticals and diagnostics. And the company that you mentioned, Genentech, is a leader and pioneer in the biotech area, which we acquired , uh, back in, I think 2009, if I'm not wrong. Quite a long time ago. And this is part of our pharmaceutical division as well. And in fact, interesting fact, which [00:01:00] I've been told and I'm eager to visit that site one day. Uh, The headquarters in the U. S. In the Bay Area is actually the same place where Genentech , uh, operations were running out of. So that's a, that's an interesting story there. Uh, Diagnostics is also a key part of uh, Roche crew , uh, and , uh, I worked for , uh, the IT team for Roche diagnostics, with the area of responsibility of business intelligence and analytics and data associated with that. Um, And , um, in the world of diagnostics, we are also a leader , uh, in terms of in vitro diagnostics, which I guess I can talk about it, then listeners are interested. Um, There are some other factors as well in terms of reach. So we are truly a global company. In fact, we are, we are celebrating 125 years since we from back in 1896 in Basel. Uh, Our headquarters , uh, global headquarters are based out of here. Um, Our employees, more than a hundred thousand plus people, are , uh, all across the globe. Um, And , uh, I think one thing which is most important [00:02:00] is the purpose of the company, which is doing now what patients need next. Such a simple purpose in such a powerful purpose , uh, which , uh, inspired all of these a hundred plus thousand people to come to work every day. Okay.

Steve Hamm: Yeah, no. I'm curious about Roche diagnostics. Does it design and sell the diagnostics equipment or just the technology for processing and analyzing the test results?

Omar Khawaja: As a patient, we all have experience of some kind of testing. And , uh, since we are , uh, in 2021, I think what we have learned in 2020 is all about go where the tests , uh, et cetera. Um, And everybody has gone through that cycle worldwide and brought this pandemic has brought all of us together in that way. And , uh, the diagnostic area is quite interesting. We somehow overlooked that , uh, and maybe , uh, under emphasize the importance of diagnostics. , uh, in, In case of Roche, we are the leader in the in vitro diagnostics, which means , uh, you know, when the test that we use the juice , uh, the blood, or , uh, urine or the tissue [00:03:00] samples coming out of human body. That's the tests that are done on those samples that could be qualified in very layman, non technical terms as in vitro diagnostics. Um, Our focus is a number of disease areas, whether it is cardiology or nervous systems or hematology or a women's health in general, or neurology. So we have tests for various things , um, In terms of our , uh, product range. Um, We have , uh, testing equipment. Uh, we for the labs, we have lab automation and software. We have disease management solutions. Uh, So both in hardware and software as you were asking.

Steve Hamm: Yeah. Yeah. Yeah Hey, I want to take the conversation up a level here for just a second and talk about just the whole field of medicine and healthcare. It seems like it's going through a real revolution with, with personalized medicine, the internet of things, making it possible to monitor people's health much more kind of intensely and continuously. Are these things creating a spike in demand for your [00:04:00] products and services? .And, And also are they creating a spike in the production of data that has to be stored, managed, and analyzed?

Omar Khawaja: That's a great question, Steve. I mean, We are all experiencing this as patients, or if you are closely , uh, associated with any life scientists for, in any, in any possible way. So I can give you examples, which I think everybody will be able to relate to. Um, We can see a big print where , uh, the healthcare and the solutions are becoming more consumerized. So, uh, Which is also driven by the various maturity levels of the digitalization of the solutions. The stakeholder engagements that we do in a classic way is changing to more digital and this was accelerated last year. But thanks to, unfortunately, thanks to the pandemic, the focus has shifted to , uh, patient outcomes uh, in various different ways. , uh, We are , uh, as a healthcare industry, not just Roche, I guess we are also , uh, focusing on how we manage this, how we measure this, how the [00:05:00] patient data is collected. Uh, It is becoming more and more real time versus a very classic approach of doing things on paper and then collecting them. As a patients aren't we all becoming more aware of these things. Uh, Look at our consumer health devices like Apple watch., But We can do things on it, which we were unable to do without visiting a lab or a doctor or a clinic a few years ago. Uh, similarly , uh, you can also now verify our own data, right? Uh, We want to own this. Um, We want to have control and visibility. As a patient, I would personally love to contribute my data for the advancements of patients, for example. Um, And then , uh, another key trend where , uh, very non-traditional players are driving, uh, we have , uh, all the big tech companies jumping in into this area. So that's a huge impact. And they're bring in with the knowledge of how to do the , uh, uh, what you can do with technology, for example. Especially when it comes to data. Big links, scalable platforms, which are integratable in the large ecosystems. And last but not the least, this [00:06:00] whole digitalization is also happening in a very classic internal value chain context, which we cannot forget. Because , uh, we are applying now , uh, more and more , uh, uh, automated process. We are applying more and more application of AI, data sciences and other processes. The awareness around data, it's quality has increased. And with all of these things happening to your point, data is being generated all over the place. I don't have a number on the top of my head which I can tell you, but I, I believe that we are generating more data than ever before, every day. And the best thing we can do with the data is to make use of it. Um, And I guess that is the direction that , uh, we are also taking in Roche. That how we take the advantage of the amazing synergies we have in pharma and diagnostics and how we uh, elevate the past of companion type of diagnostics, where we used to have medicine for every other [00:07:00] solution, , uh, moving towards a more targeted medicine for a different disease areas and going towards this individualized treatments, which will only be powered if we are able to make the sense of the data and apply advanced analytics on it. So that's the future. It looks like with all these trends taking place.

Steve Hamm: It's really interesting. You know, when I think about it, you know, I think in the past, a lot of that real time dog diagnostics data, or, you know, near real time stuff was all just like a single, you know, a single line of data coming in, but now it's, it's integrated. You, You might have five different kinds of data that addresses a singular , um, issue or something that is happening with a patient. But it seems like, you know, having the data all in one place and shareable and integratable is like absolutely critical to the, to these, the revolution that's going on in, in medicine today.

Steve Hamm: Now you mentioned before the COVID-19 pandemic, Roche Diagnostics has been in the middle of [00:08:00] this crisis. You know, You have a bunch of tests, the molecular, the antigen, the antibody tests, each kind of with its own strength and weakness or purpose. So if you could tell us, how did Roche react when the crisis emerged and how has it responded to the pandemic since then?

Omar Khawaja: Yeah, This has impacted everybody in a diff in a mega scale, right. worldwide. We are impacted personally and also as a professional level. , um, as a Roche , uh, if I can try to summarize, I think, As a leader in the healthcare industry, especially in the diagnostics area, we have an integral role to play here, and that's what we have done um, you know, with the availability of all the tests or the different type of tests that you have mentioned , uh, We have made sure that , uh, the testing solutions , uh, in all different shapes and forms, they are available all across the board, in all the geographies where they are approved. And we have been working closely with our partners and governments, healthcare providers, and other various players in the industry to make it work. Plus we have not to lose sight on what's happening outside the COVID world as well, with the remainder of the [00:09:00] portfolio commitments that we have. So we need to also make sure that we are able to provide, continue to provide those products out there or whether it is tests or medicine. So, um, internally if you speak. we need to take care of the employees as well um, and the partners that we work with, so how to make sure that we continue to work in a virtual world , uh, that have an in a safe environment where it was necessary still to come to office. And that has been on top of our agenda since this pandemic started. Um, Or at least since I joined Roche last year in the middle of pandemic and experience some of these , um, amazing practices that are in place, which help us operate in this pandemic times.

Steve Hamm: Yeah. Now you had these diagnostic tools and protocols and all that kind of stuff. What has the role of data and data analytics been in your response to COVID-19?

Omar Khawaja: Oh boy. Um, in so many different ways where to start even, right? So, um, you one can imagine , uh, that , uh, the impact of such pandemic and the [00:10:00] demand of the tests, whether COVID related or non COVID related and how they impact will be with all the travel restrictions taking place, what will be the impact on the supply chain of the goods, but that it's for manufacturing or for distribution. And in all of those challenging situations, data has to play a key role. And that's what , um, how to map, how to use the data, to manage the you know, an absolute increase in the demand overall and how to make sure that our courts are available. We, there, We make the best use of the data. Uh, Also applied some advanced , uh, distribution and forecasting algorithms that the data science teams have done. I'm really proud of the team who did that. Uh, We looked at how , uh, what will be the impact, for example, when it comes to altering, what will be the impact on , uh, uh, HR , uh, related topics on employees and , uh, another aspect where data beta keyboard was that it was very clear that , uh, internal data is not enough. We need to also rely on a lot of external data, for example, some of it, which is now easily available on the [00:11:00] data cloud , uh, like the star schema data sets that are available worldwide. Numb. Those, Those were key. And in the early days , uh, it has been a challenge on how to get that, how to make the best use of it, and then deploy it in various different user scenarios. We also had a very interesting learning related to data and analytics that, what are the gaps? And that kind of similar to the acceleration of the digital revolution. I would say. , uh, also accelerated our learning towards that journey of becoming a better data argument. Uh, We learned how , uh, we are sharing the data or what are the challenges in it? We learned about how we need to have , uh, uh, infrastructure that supports data sharing in of course, in an easy and govern access. Uh, We also learned that in order to make decisions, we need to empower the decision makers. So the freshness of the data, the insights that are generated on the timely manner, becoming more and more important. So all of these aspects are, have had a huge impact on what we have done with the data during this time.

Steve Hamm: It's interesting when we [00:12:00] have crises or big stresses on our systems, we really spot the flaws in the system, or the cracks or the gaps. And , uh, and that's when you have to learn the lessons to, you know, how to be ready for the next one. So it sounds like you've been doing that. Now, almost everybody we talk to on the podcast talks about some major IT or data transformation that they're undergoing. It seems like this is perpetual. You know, In fact, I think when I think back on my life covering technology, I think ever since about when I, when I started doing it in 89, transformation was on everybody's lips and it still is. So I guess it's just a continuous process. Uh, If you would talk about what you're doing with IT and data transformation per diagnostics and, and you know, more broadly for the whole company,

Omar Khawaja: So, um, As, as I mentioned during my introduction, I'm part of the Roche Diagnostics IT team. , uh, I'm leading the BI and analytics team. And you're right. Uh, the transformation has become , uh, not just a buzzword, it's a reality. Uh, However, I think the let's, if you simplify this, the [00:13:00] reality is that the change is the only constant. So, how do we stay edge line? How do we stay on top of things by learning new things and adopt and adjust? That's what to me personally, and that's how I , uh, treat the transformation. It's It's about learning and growing at the end of the day. Uh, so, uh, We are of course going through various aspects of it , uh, learning , uh, in terms of how the organization is set up or when it comes to technologies or when it comes to the architecture, how to apply the modern ways of working, how to live in this ecosystem. And that's the approach that we are taking, at least in the, in my team, in the BI and analytics area.

Steve Hamm: Yeah, no, I understand that Roche overall, the two, the two big divisions and headquarters, I guess two very decentralized. What I would think that that would cause kind of challenges, but also create opportunities , um, you know, for, for people using IT, sharing data, all that kind of stuff. How do you kind of overcome the [00:14:00] challenges and take advantage of the opportunities?

Omar Khawaja: that's, That's very true. One of the key aspects I've learned and experience in my limited time over here is that that Roche has an amazing culture of decentralization, which leads to empowerment to the teams on the ground who need to make things happen. I love it. Um, It comes with its own challenges. Uh, I think both centralization and decentralization , uh, the extremes of both will result in challenges and , uh, advantages. So in my personal view, , uh, how can we drive the balance between the two? There are certainly a few things which make sense to centralize and make sense to govern. And there are certain things that they are best left done when people are empowered, people are able to move fast, make decisions. And how do we bring these both worlds together? And unfortunately, it's very simple to say that. Uh, But the reality is that it's very difficult to be simple you know, in this. And that's , uh, that's the challenge. Um, uh, I do see a lot of things that we can do, especially when it comes to [00:15:00] data. Uh, Imagine the teams who are responsible for the data, which are decentralized. They are empowered. They feel empowered for the end-to-end life cycle management, rather than some central team setting completely disconnected. And that's a big advantage that we can leverage. Um, and , uh, I, I do see a huge potential and the huge model , uh, that this decentralization can help us over there.

Steve Hamm: I'm just curious how you make this work. Do kind of all the divisions share some standards for data management? Do they all share some basic platform technologies, but then each of them can do their own applications on top of it, or is it more complicated than that?

Omar Khawaja: In the short answer is that it is more complicated. Uh, The long answer is that it depends on , uh, how you look at the whole, let's say, infrastructure stack behind your typical data and analytics solutions. It makes a lot of sense to centralize things when we [00:16:00] are talking about the crown level infrastructure, whether on-premise or cloud, how you manage that. Um, I don't think there is a need to uh, decentralized that part. We don't have to reinvent the wheel all the time , uh, When it comes to , uh, taking advantage, if you look at the other side of that spectrum, let's say end of the spectrum over here, , um, we do have certain needs not everybody's, a for example, is interested in unstructured data for us, for instance. It is a need for a certain use cases, and some departments and functions are quite , um, uh, knee-deep in the area of unstructured data. So would it be better that that team is responsible and we make them responsible and accountable for the best solution when it comes to unstructured data? And that's the approach we are trying to take. Um, There is a lot of things that have happened in past how the, how the whole journey started with the company. Um, And what our direction is that centralized where it makes sense. And given proper empowerment where it's needed to have speed, agility, and make an [00:17:00] impact on the customers and patients at the end of the day.  That's, that's our objective. We have to put them in the center of our decision making and not about, Hey, who manages the platform?

Steve Hamm: Yeah, I would think in a situation like you're in , um, yeah, I would think that in a situation like the one you're in, leadership, collaboration, even diplomacy, are some of the key things that you need in kind of your, your portfolio, your management toolkit. So you've worked for a number of companies in your career. I wonder what are the most important lessons you've learned along the way and how are you applying them now at Roche?

Omar Khawaja: Oh boy , uh, did absolutely right. Uh, My career started long time ago. Um, And I'm grateful to so many people who have, have a great influence from the day one, even when I was studying on what I have learned, what I become. Whether it's my, during my personal life or professional life. I think that it comes to uh, the leadership , uh, when it comes to applying those lessons. , um, It is, It's so important to remember that when I start, for [00:18:00] example, when I started at the very humble beginning in the software development area, how to , uh, you know, uh, apply those best software engineering practices. Now, how they really relate to leadership I can totally see that I can have a very good, meaningful conversation with my data engineers. Um, At the same time, I've been lucky, really blessed to work with great leaders who taught me the values of passion and patience and perseverance. And , uh, uh, they, I remember one of the leaders once told me "Excellence has no boundaries." And that just stick to me. Um, in, In general, I believe that one should have a can-do attitude. As a leader, we are responsible to serve our teams, , uh, have a humble approach of how we can remove the roadblocks for them. And that has helped a lot, even when I joined Roche last year. Um, And it was right in the middle of pandemic. So imagining, imagine waking up on your day one and stay sitting at a desk in your home, and you are on your new job. Um, but , uh, Overcoming all the challenges, which were absolutely [00:19:00] nicely done , uh, during my onboarding, which was totally planned. And how to make a connection, even in this virtual world , um, that those lessons have come really helpful for me. Um, And I can only , um, encourage everybody, I even wrote about this, that one should have a can-do attitudes toward these things. I can totally understand that not everybody is in a situation where you can work in a complete virtual environment, but yes, when there is a will, there is a way, and , uh, you can do things if you are willing and you are willing to adjust and learn. And that, to me , uh, this learning agility , uh, I would say, has come a long way in my career.

Steve Hamm: It's interesting that I read just today that Snowflake no longer has an official headquarters in California. Its headquarters is in Bozeman, Bozeman, Montana. So that's just a real great example of a company that's recognizing kind of the virtuality of its existence. And , uh, I thought that was pretty interesting. It really [00:20:00] is it's kind of a, it's a symbol of a big change that's happened in the world for business and for people. So it's interesting. Now you mentioned you, you came on about a year ago or last year. So I want to ask you a little bit about the history of Roche Diagnostics and cloud and moving data to the cloud. I know it's before your time, but if you could kind of fill us in on that, how did it get started and what challenges did the company encounter and how did it overcome them?

Omar Khawaja: that's, That's very true, Steve. So, um, the journey already started before I even joined and it , uh, we had number of pilots, POVs, projects, you name however you want to name them, , uh, that were running in pockets. Uh, but in the area of , uh, I think , um, before we jump into the data and analytics part, But , uh, generally people to start with even some transactional type of application, then the journey of using the cloud in that perspective. And that kind of opens up the conversation even for data eventually. Um, and , uh, Similar journey must have taken place in Roche before my time. Um, But when I joined , um, one of the key tasks was to look at where we are, [00:21:00] have a strategic view of things. And that's what we did towards the quarter three last year. And we took a strategic decisions on how our approach will be in building this , um, data and analytics solutions platforms for our business and customers and patients. And that's where , uh, we had a more strategic focus on how we will approach. Uh, our setup on the cloud. Um, I think in terms of challenges, you name it, we have faced it, even in my short time over here. , uh, And I can relate to even my previous job that I can definitely , um, Uh, I have faced those challenges. Uh, I may have even asked those questions in my own personal journey. Do it , uh, To start with , uh, I will repeat what I, I guess everybody says, the mindset. Um, We need to open up, we need to be open-minded. We need to really understand what does it mean to have a cloud first approach towards things, that this starts with the mindset. No, Nothing else. Uh, if you don't have that, I think that will become the roadblock. And no matter what we are doing , uh, the architecture decisions and the approach we [00:22:00] need to have completely changes. So we need to be aware of this. We need to be open. We need to be , uh, open to learn how to do that. The way we approach building things, it's very different from what you do on premise. Uh, Gone are the days when you are ordering a physical server, waiting for the delivery, installing it in a data center and then configuring it. I mean, those things are still important, but maybe not for company like us. They are important for company another company, perhaps. Uh, And I've done that myself. I I've built data centers. I've experienced those weekend nights where we said, oh God, please, so we have just gone live. Please keep working over the weekend. For example, , uh, the ways of working changes you, you , uh, you need to adopt to that. Uh, One of the aspects is costs difference between the capitalization and pay-as-you-go model. It's a very different mindset. It's not about, oh, can I have this compute power or X amount of storage? It's about, yeah, do you really need it? Then you can have it. [00:23:00] So real focus has transferred from the limitations of a typical infrastructure to a opportunities that the cloud provides, shifting your goal towards what outcomes are we expecting. And last, and absolutely not least, and maybe second, most important thing is learning the security aspects of it. Uh, We need to understand in depth how the security works. We need to ensure our data is secure. It's our responsibility. And as an IT leader, as in , uh, in the area of data and analytics, I would say that is the core thing that I must ensure, , uh, having two partners to work with who will also provide solutions, which are secure. We have a responsibility in general to the company to make , uh, the cloud journey, not just for the sake of a cloud journey, but also to have those outcomes in a secure and a compliant manner.

Steve Hamm: That's interesting. [00:24:00] Your partners, the health care providers, and the clinic, you know, the insurance companies, they're the ones with the patient data. That's kind of puts it all together. You, You have these discrete pieces of it, but it still has to be secured, you know, under all sorts of regulations by countries and the EU and things like that. Right?

Omar Khawaja: Exactly. And we have to respect them. We have to respect those boundaries. We need to be compliant to them. And those are the things that have to be at the center of our journey and the strategy that we are setting up. Um, that's, That's a key point, Steve. I'm grateful to one of my colleagues in U.S. and my previous company who actually introduced me to Snowflake. , uh, And , uh, I'm we were in our typical journey of on-premise, doing things on on-premise databases and data warehouses, , uh, shifting to data lakes, using Hadoop and other such things file-based system approach and SDFS, whatnot. Um, again, a massive learning for myself. I'm no expert there. Um, Having said that we were looking into how we take the next step. How we can [00:25:00] scale up , Uh, from , uh, where to go from there. Uh, Businesses, business teams were becoming more and more data savvy. I would call them the citizen data roles are becoming a norm now. And Which means that IT also needs to change its approach, how we work with them. Uh, Another key learning for me, I would say. And that discussion and those changing dynamics and the external and internal environments have led us to look into what would be the right. Or what would be the options to go ahead? And that's where we started our journey, comparing Snowflake with other solutions that we were using and some of them which will be, we're not using. We did a massive thorough exercise of comparing things, using different metrics, research, ensuring that the IT securities and compliance controls are met inside out.

Steve Hamm: But Omar, was this at your previous company or is this a 

Omar Khawaja: So this was, my journey started in 2019 on this. So that was my previous company.

Steve Hamm: I got you. I understand now. Yeah.

Omar Khawaja: And at [00:26:00] the same time when I joined Roche, guess what? Another team was in a similar boat. And I'm pretty sure you guys also face that in all of the companies. Uh, But we joined hands , uh, and I guess the outcome was similar to what I've experienced firsthand. And now. Uh, it is one of our key capabilities as part of our data and analytics platform, capability capabilities offering.

Steve Hamm: Yeah. Yeah, he, Could you describe a couple of the key applications that you've developed , uh, using the Snowflake platform, the data cloud? And talk a little bit about the benefits you achieve with it?

Omar Khawaja: Sure. Um, Our journey in diagnostics and my team, or Snowflake , uh, is a relatively fresh, but very promising, so far. Very good experience response. We are aspiring to have our entire on-premise uh, solutions and different shapes and forms migrated to snowflake eventually, which runs across the internal value chain of the business. Um, which , what, What does this mean? That there is a lot of excitement everywhere, but [00:27:00] we need to start somewhere , uh, and we can not do everything at the same time. Not at least at the very beginning or not at least last year in quarter four, quarter three in 2020. So we started , uh, with the area and what better area to start with with customer experience and insights. And that was our first project, which is went live last year in just eight weeks just to let you know. , uh, That was unheard of. And we delivered that together with a very large , uh, not very large, but a very good , uh, and competent team , uh, who were brave enough to take the first step. Thank you to those people. And at the same time , uh, by the way that team is now on the fourth release of their product, product and project, however you want to put it. Uh, But we have number of things now running since then. We have , uh, initiatives going across manufacturing and quality and regulatory and supply chain. There are a number of VOC have already started in other functions as well. So we aim to , uh, multiply the efforts [00:28:00] now across the different functions, because we have now the opportunity to really scale. That That limitation is gone. Uh, What we have learned in the first , uh, months of implementation is that how to enable those teams better. , uh, Whether it's about skills or processes, how the ecosystem of tools needs to work to make, make this journey possible. And that has really helped us , um, that how we are collaborating with a larger IT teams, as well as with our business colleagues and other IT functions to have this , uh, parallelism, I would call it in place so that we can operate across the entire value chain and not just in one function.

Steve Hamm: Yeah, that's interesting. It occurs to me that, you know, one of the big things that's happened here is that in the past a lot of, you know, software development, project development, was a lot of it was about dealing with limitations in the system. And people now have to kind of like, turn it on its head and say, no, we don't have limitations. What will we do with that capability? And I would imagine that for people who were, who were trained and, and all their [00:29:00] experience in the old system, it must be quite a change, a shock almost, to , uh, to deal with the new environment that they've got.

Omar Khawaja: That's true. And that was my first comment in terms of learning, right. That you have to overcome the mindset, the Approach needs to change. And , uh, we, we cannot ignore that. Uh, that's a very important aspect to keep in mind during this journey.

Steve Hamm: Yeah. Yeah, Now, I understand that you're using the Snowflake technology to create a data mesh for Roche diagnostics. You know, I've heard the term data mesh. I've also heard data fabric and I I'm, I'm not sure at this point, you know, are they the same thing? And what what's the data mesh? If you could start by describing that and then, and then how you're using it, that'd be, that'd be great.

Omar Khawaja: Oh boy. Is it a buzzword? Is it the next buzzword? I don't know. I can not comment on that aspect, but I can definitely share my own personal point of view and , uh, after implementing all shapes and forms of data warehouses and data marts on different technologies, , um, ranging from a very classic approach with a very limited , uh, impact [00:30:00] and scope of work to a global deployment, creating data lakes, creating enterprise data warehouses. I've done that all. And I'm lucky that I've done that all I've learned a lot and. Those technologies, those architectural paradigms and practices have delivered what they will promise to deliver. And as it goes in life, all of those things also have their limitations, but whether limitation was technology or not, that was not always the case. It was a key driver and a key aspect. But , um, uh, in, in 2019, if I'm not wrong, The creator of data mesh Zhamak Dehghani, , uh, who works for ThoughtWorks , um, when she wrote her first article, um, I'm not, I'm not kidding , um, when I read the article, I was like, Hey, the first half got an article really summarizes my entire career. And at the end of that middle of the article, it says that that is not working. And I was like, what have I done? And I read that article six, seven, maybe eight [00:31:00] times, different angles, different times of the day to make sense of it. And I'm. I am very grateful to Zhamak who I then approached taking the advantage of COVID-19. I said, Hey, hi, I'm Omar. I would like to learn what you have written. You have turned my world upside down. And I must talk to you. She was very kind to speak to me , uh, and helped me understand where she's coming from. And now she, there is a lot of literature and practices out there and, and. I don't know whether I'm qualified to explain to you, but I can explain my own view that it is addressing. Not just the technology part, but the four key principles around data. It talks about , uh, not having these centralized team to do everything, but having a more de-centralized domain oriented approach. And if you can link back to my earlier comment on the culture that we have, that fits very nicely in our culture of having [00:32:00] decentralized ownership of the data. And that's what is uh, one of the reasons that attracted me towards the whole data mesh approach. , uh, The other aspects. Are instead of doing these typical projects and then throwing the ball over the operation boundary line to somebody sitting somewhere in the world, managing this application, that does not work anymore. Um And we're just talking about the DevOps piece of it. I'm really talking about the ownership of the data as a product. And that is one of the principles of data mesh as well. Um, The third aspect is around self-service data and analytics platform where technologies like Snowflake potentially comes in. And that platform approach is also done in a way that you are treating your platform as a product, taking inputs from these decentralized teams on what capabilities and functionalities that they need to create those amazing data products, which needs to be used and reused and reused across the enterprise. Um, And the last aspect is [00:33:00] governance. Uh, When we are talking about such autonomy, , uh, such freedom, we need to have also some kind of guard rails. Uh, It can be in the shape of a prioritization, portfolio management, data quality policies, data governance policies, data architecture practices, et cetera. And , um, that's why this approach is so uh, different from our classic approaches of, Hey, let's build a data lake and dump everything somewhere. And then we will decide in, after we have spent months and months of dumping the data what we will do with it. With a data product approach, the first and foremost thing that you ask is that who's going to use my product? What are they going to do with it? And then you talk about data. Changes your approach towards it. And that is nothing to do with technology ironically. Um, And to really conclude , uh, in my point of view on this topic, Steve, is that what Einstein said long time ago, I guess. It is you know, insanity to do the same things and expect different results. So we have [00:34:00] faced data challenges on data lakes and data warehouses, and this approach is different. It is talking about various categories of things that we are doing differently. And that's why I see it very promising.

Steve Hamm: Yeah, hey, say the name of the innovator for data mesh again.

Omar Khawaja: Uh, her name is Zhamak Dheghani.

Steve Hamm: Okay. Yeah, I think we should invite her to our , uh, to the podcast to be a guest. I think that would be a really fascinating episode. So.

Omar Khawaja: Oh, plus one, too, that I will let her know.

Steve Hamm: Now, one of the things about the cloud is that it really enables powerful data sharing. And I wanted to hear a little bit from you about how Roche Diagnostics is using data sharing, or how do you see using it in the future?

Omar Khawaja: Data sharing, it is such an important learning. And if I, if you relate to what I was telling you about the COVID 19 story, it , uh, it really reinforced the importance of data sharing. Um, it also, I see three flavors of it, Steve. Um, I see that we need to share data within Roche, within [00:35:00] functions, within diagnostics, between diagnostics and pharma, where appropriate. Uh, I see a potential of sharing the data with our ecosystem. One aspect of that ecosystem is with our customers. With them. Uh, And the other aspect is of course, with our suppliers and vendors, for example. , uh, The third aspect is of course, the other way around. We don't work in isolation. We work with a lot of companies who provide us the data. And in the life sciences industry, I think I can safely say the name. We also have , uh, other , uh, market research agencies working, doing primary market research, et cetera. And guess what? The end result of that is data. And instead of , uh, spending time in data acquisition, , uh, that data sharing , uh, that has now possible through as simple as data marketplace approach. Um, I see a huge potential there and a huge usage for us. Um, We can get the access to the data that we need without spending every month or every week or every day data [00:36:00] acquisition uh, type of activities, whether it is resources , um, computer or people or money, or , uh, we can avoid duplicating the data. We can reuse the data on its own instead of , uh, making copies of it so that others can use it. Um, That's why I love the uh, segregation of storage and compute concept. Uh, Those of you who are familiar with Snowflake, I hope you can fully appreciate that , uh, uh, concept and , um, the speed gains. At the end of the day, the customer needs to make decision internal or external. Can I generate insights on time or will I spend time to get data loaded in my data warehouse in four days? We can get rid of that phenomenon, using the data sharing. Capability and have access to the data. The moment it is shared with us, generate timely insights. That speed game is tremendous, and that promise we, of course, we need to also practices in a number of ways. I must admit. Uh, But the [00:37:00] technology is real. It works. We have seen it. We have experienced it. Uh, and, And that's a great functionality to have when it comes to data sharing.

Steve Hamm: Just from our conversation so far. I mean, it's clear that, you know, this whole field of , uh, you know, data engineering , uh, business intelligence, all of these related areas has been changing incredibly rapidly in the past couple of years, but I'm going to ask you to , uh, to look out now, like a year or more, just, you know, relatively short term, what do you expect would be the major trends in data management and data analytics?

Omar Khawaja: oh boy. Um, I don't have a magic , uh, ball with me, but let's see. Um, I think , uh, people will realize that they have to work with the ecosystem of tools, but it will get more and more integrated eventually so that we don't have to jump from tool A to tool B to tool C. And that's happening rapidly right now. Uh, We will also see the interoperability between these solutions will increase. Take the example [00:38:00] of recent announcements on external functions that you have. That's a classic example. Uh, I foresee that the large amount of work that we do no matter where you look at, it's still pretty much with structured data. In some cases, semi-structured data with XMLs and whatnot, and the value that you have in terms of JSON files. But I foresee that things will move towards text analytics and voiced based images are completely into the area of one structure, basically. Those The value in those types of data sets can be unlocked. And very soon , um, I do see acceleration around data science and application of AI, definitely. Um, I'm hoping and relatively, even. Pushing forward, to make things more and more , uh, productionized and industrialized in the hands of the people who need to use AI and data science and , uh, Go beyond the proof of value cycle and the milestone that we have achieved. So these are the top two, [00:39:00] three things I see that will be happening very soon in the next year or so.

Steve Hamm: an amazing lineup. And it sounds like you think AI is really going to be democratized. I mean, really put in the hands of all sorts of people, you know, even business people. Right.

Omar Khawaja: Yep. Absolutely. I mean, they have to use it, right? So if a sales rep or a service rep needs to make a decision and AI is needs to help them make that informed decision, , um, what good that AI will go to them is in the hand of a data scientists. It is good in the hands of a person who will make use of it .The customer of that algorithm, the end user of that algorithm. And that will increase in my point of view. And that needs to increase if you  really want to make an impact.

Steve Hamm: Yeah. Well, you can imagine a time when the end user doesn't even know they're using AI. It's just so embedded in an easy to use, that, that kind of thing.

Omar Khawaja: That's the future, right?

Steve Hamm: Well, let's talk about the future as well. I mean, you've talked it out a year and you really packed a lot [00:40:00] into that year. But now I'm going to ask you to put on your visionary cap, let's look out five years or, or even more. How do you expect the data cloud to impact business and society? I mean, beyond, beyond your industry, just the whole thing.

Omar Khawaja: So let's, let's take that a full back one year and explode it. Right? So, um, data sharing will become a norm. People will laugh at the old ways of dumping and CDs and uh, data sharing via FTP and whatnot. That will be like, are you kidding me? Who did that? Uh, You know, it's the same way our kids react to a landline phone? Uh, my, My son, 10 years old, have no idea how to talk on the phone because they are not used to it speaking. They, They may not even understand when we make the sign of a phone with our hand you know with a finger and a thumb.. That's, That's the reality. That's the, It's funny, but it's a reality. And similar things will happen when it comes to data sharing. We will say, yeah, just publish it and I will access it. Yeah. Thank you so much. That's it, really. Are the, Did you tell the journey that COVID [00:41:00] has started? I think that will explode. We know it's important. Let's not shy away now. We have proven it. The world knows it. But that also means that the data explosion will take place. All type of data sets. Uh, The new normal will means that we will no longer be working every time, face to face. So maybe , uh, the advancements in telemedicine and things like that, will explode. And that, And the tools and the technologies and the understanding of the non-structured data, will explode big time. Uh, It will be as easy as writing an SQL statement in future. Uh, I think the, on the data sharing piece as well, the companies will realize that they need to work with each other. , the, The chain world will I think we have already done this this year, last year. Uh, The competition moved on into more coopertition and the favor of let's say consumers in the favor of patients in our case. And , uh, to build upon what you were saying about AI, it has to be in the, it will become a given. It will be, We'll take it for granted. It will become so normal. Um, And , uh, we will ask [00:42:00] ourselves the question, where was it? How. How about people doing things, right? So when I start writing something on my email, some email clients literally complete your sentences based on your writing styles. Um, And it can be, it can be, and sometimes it is so helpful because that's exactly what I wanted to write and boom, I'm done. So I think that will a similar thing will happen when it comes to data. Uh, And this will have a tremendous impact on the business in general in society , uh, um, for the, for their betterment, that how we will be using data out there.

Steve Hamm: Yeah. Yeah. You know, Omar, we typically finish with kind of a personal fun question and I wanted to know , uh, you know, so what do you do when you're not working? Do you have some hobbies or a passion that you can share with us?

Omar Khawaja: These days the really the hobbies are been tremendously impacted, but , uh, I have , uh, a little , uh, quirkiness in terms of a habit that I've developed over the years. Um, I collect a lot of things. So the secret is out. Uh, I have a [00:43:00] huge collection of , uh, mugs and shot glasses all over the world. Hundreds of them. The downside is that I have to take care of them, sometimes test them. And since I'm at home, that's a huge responsibility. Uh, But there have been a pause on that , uh, all activity because of pandemic and I'm really looking forward to again starting it and , uh, collecting those things again. I'm quite passionate about it. Um, I have , uh, called out my taxis and asked my friends to take me to a specialist stores many times just to do that. So that's, I find it funny, maybe not for others.

Steve Hamm: That sounds like a lot of fun. Yeah. I used to travel a lot with, I was with IBM for a while and Business Week and my son, much younger then, would always say insist that I get him a shot glass and I would get, I would get him the tackiest shot glasses I could from wherever I was. And he really loved them. So it is a fun thing and it wouldn't it be great to travel again.

Omar Khawaja: Yes. I think the, I would love to do that, uh, love to meet my team. I've not met anybody in my team [00:44:00] in past 14 months. I've been completely operating virtually. Um, I'm keen to do that in a safe and healthy manner, of course. I hope we will not go back to that travel extreme situations, travel fatigue, I guess. Uh, so, uh, maybe, but there has to be some. Yeah,

Steve Hamm: Yeah. Yeah. We're going to find a happy medium there, I think.

Omar Khawaja: I guess so.

Steve Hamm: It's been great talking to you today. I think, you know, You've really shared some wonderful insights, not just about like technology, but also about leadership and how to get things done. And , uh, and I, you know, it, it really came through to me very strongly, how this COVID-19 pandemic has really kind of forced a lot of changes, an acceleration of changes in the way we do business and the way we do IT that is a positive. So at least something good is coming out of this. And that's , um, a bit of a relief since it's been, it's been a very tough time for the world. So thank you so much.

Omar Khawaja: Very nice talking to you, Steve.