Episode

13

Season

1

How AI and Robotics are Transforming Industry 4.0

With
Andrey Shtylenko
Global AI Solutions Architect at Honeywell
Listen this episode on your favorite platform!
Listen in browser!

Podcast Questions

Transcript

Joe Toste (00:15):

You're listening to the TechTables podcast, a weekly Q&A podcast dedicated to interviewing industry leaders from across the world, ranging from startups to Fortune 500 companies, mixing it up each week with topics ranging from design and product innovation to IoT and Industry 4.0. Let's do this.

Joe Toste (00:33):

Hey, guys. Welcome back for another week in the world of TechTables with me, Joe Toste. [inaudible 00:00:37] behind the scenes on LinkedIn, Twitter, and Instagram. There, you can even message me questions for future guests coming on the show. But today, I'm super excited.

Joe Toste (00:44):

We're going to shift our focus to AI and robotics with Andrey from Honeywell, as we chat about his previous life as a successful startup CEO with three successful exits, helping Honeywell on its journey from industrial conglomerate to a software industrial company, the importance of commercializing quickly, the extreme demands of e-commerce and logistics on warehouses today, helping companies eliminate the risk of injury with the connected worker movement, how AI brings new paradigms, how engineering teams can adapt a data-driven mindset, and the critical importance of focusing on the customer problem before diving into POCs and full scale, and, of course, Andrey's three favorite books right now.

Joe Toste (01:21):

Huge thank you to Andrey for taking time to come on with me today, but that's quite enough from me. Without further ado, I'm thrilled to welcome Andrey, Global AI and Robotics Leader at Honeywell. All right, Andrey. Well, thank you for coming on the podcast today. Super stoked we were finally able to make this happen.

Andrey Shtylenko (01:36):

Hey, Joe. Thanks for inviting me.

Joe Toste (01:38):

Love it. Okay, so let's kick off today a little bit about yourself and your entrepreneurial background over the years launching several startups, and now leading a team of data scientists and machine learning engineers at Honeywell SPS. Let's start there.

Andrey Shtylenko (01:51):

Oh, sure. Absolutely. If you ask people what they know about Honeywell, the majority of them will probably say "thermostats." And this might seem like an actor who is famous for their one role, but with Honeywell, it's much more than that. [inaudible 00:02:06] that 90% of large air carriers, 25% of commercial buildings, including Burj Khalifa in Dubai.

Andrey Shtylenko (02:13):

We produce safety equipment for a half a billion people, industrial workers and first responders. Also, we have one of the largest install base for industrial waste systems. We are one of the top three world's major manufacturers [inaudible 00:00:02:27]. So all in all, that's a pretty interesting reason to wake up in the morning being very motivated in working for this company. So I, myself, facilitate to the analytics engineering efforts across lines of businesses within safety of productivity business group.

Andrey Shtylenko (02:44):

On the strategic side, I work closely with our global CTO to develop and execute our AI strategy that has, of course, multiple facets, from talent and capabilities, infrastructure partners and vendors, processes, and organizational alignment, and on the tactical side, manage a team of machine learning engineers and data engineers who work for respective business lines. So my team works on developing [inaudible 00:03:08] models for a broad variety of portfolios, from barcode scanners, smart safety equipment, sensing equipment, gas sensors, besides many other things, of course, like warehouse automation and robotics for warehouse. So that's a good overview of what I do at Honeywell.

Joe Toste (03:26):

Oh, that's great. And how many directly are on your team right now?

Andrey Shtylenko (03:30):

So I have roughly a dozen machine learning engineers and scientists, and then have many more who are metrics reporting to me and that I'm also organizing their work.

Joe Toste (03:44):

That's awesome. And now are you in the weeds still, too? Are you coding still or are you more assembling division, the team managing it?

Andrey Shtylenko (03:52):

Oh yeah, I'm full hands-on, but I wish I had more time for that. I see a lot more potential actually organizing the work of engineers rather than coding myself. But that's what I do in the evenings when my family is asleep.

Joe Toste (04:06):

Love it. I do want to chat about Honeywell CEO Darius, who believes that the Cloud, big data, and artificial intelligence is going to transform Honeywell from an industrial conglomerate to more of a software company. Where do you see your team playing a role in that digital transformation journey?

Andrey Shtylenko (04:24):

So yeah, we have been perceived as a thermostat company, but at the same time, we're moving towards being a software industrial company. And what that means is we already have a huge install base for our solution, but at the same time, we need to focus on the software solutions. And this is where we invest a lot into initiatives like our own cloud platform on which all the other software solutions are based on. We also have investments into our own robotics platform that right now we're intensively using. All in all, it's not that we never had the software experience, but we are refocusing our approach to selling our solutions to have it software-first, of course, and moving away from one-time sale to recurring revenues.

Joe Toste (05:10):

Yeah, no, a hundred percent, it does. And I think, where I was reading this... One of the things I love is investments and investing, and I actually was reading this article in Barron's maybe a couple of months ago, and it was really good. It was talking about this transformation that Honeywell was going through, and they'd interviewed Darius. And as they were, he was really positioning that the company in that cloud/big data/artificial intelligence, also to change that perception that you mentioned, right? And it's hard to change that perception because Honeywell, I don't know, it's been around for a hundred years or something. But it was a really, really great article just as far as the transformation that was happening in so many facets that you mentioned, from the cloud platform to robotics, not just thermostats. And although I do have a Honeywell thermostat in my house, which is really funny, and I can control it with an app on my phone, but yeah, it's much more than that.

Andrey Shtylenko (06:03):

The funniest thing is actually that we no longer produce Honeywell thermostats. So it's a spin-off company called Resideo, which is actually producing Honeywell thermostats with the brand of Honeywell.

Joe Toste (06:15):

Oh, I'm so disappointed now. That's not the same. The Honeywell thermostat in my house, I'm going to look at it and say, "It's not really Honeywell anymore." So there's a lot of shiny objects in AI and ML that can distract leadership teams and engineers from actually solving business problems and outcomes. When we look at the intersection of Honeywell SPS and AI, how do you decide which problems to solve and what work matters most?

Andrey Shtylenko (06:39):

Sure. Great question. So, first of all, besides the basic things like outlining the size of the opportunity and outlining the strategy, I personally go through a couple of points. First of all, is of course, does the application actually solve the problem? So the many ideas that are coming through that are rather about testing technology hypothesis rather than about solving a customer problem. So when we are evaluating different ideas, the first question is: "Is this actually a product which we are going to be commercializing right away, or this just a R&T [inaudible 00:07:14]?"

Andrey Shtylenko (07:15):

Another thing is probably looking into the aspect of, "Can we convert this opportunity into recurring revenue opportunities?" So if that aligns with our software strategy... Going away from just one-time sale into something that becomes a service. Another thing: How quickly we can create a prototype for data collection and preliminary inference, and be able to deploy [inaudible 00:07:38] our current install base as soon as possible. And maybe this prototype is not going to provide the full decision-making support. Maybe it's going to only show some confidence scores, maybe it's going to be having a human in the loop, but we are big fans of deploying fast end-to-end prototypes that users can start evaluating right from the get-go. And if we are able to do that, that's definitely a plus.

Andrey Shtylenko (08:03):

And then, of course, whether we can create a competitive advantage based on the patentable IP or the amount of data that we can collect [inaudible 00:08:11]. So I do believe that the algorithms are becoming more and more commoditized, and so competitive advantage is created more from the data that you are able to collect to train those algorithms. I think this is a very important competitive advantage and it's a very important factor. And then, the final one probably is what is the cost of the error and whether there could be a human-in-the-loop situation, and at what price points? So if the cost of error of the AI solution is too high, and then there is no way to include, for us, a human-in-the-loop, we are probably unlikely to precede that right away.

Joe Toste (08:47):

That's really great. One of the things you mentioned, I was just jotting some notes down, was competitive advantage. When you think about that and where Honeywell's at and along with... You sold a couple of companies, you've had some really great successful exits, how are you helping to position as much as you can within Honeywell to build that competitive advantage for Honeywell out?

Andrey Shtylenko (09:11):

This, of course, relates to being able to see where can we source those massive datasets from the current install base, and how can we use those datasets in order to develop even greater customer benefits, customer value. I'm coming from the perspective as a consultant. My previous entrepreneurship experience was actually coming up and seeing whether we have those pockets of data and whether we are able to deploy those to create new machine learning models and to add additional benefits to the customer.

Joe Toste (09:46):

Yeah, I think your previous background running with the entrepreneurial, running a couple of startups, successful exits, and even failing, taking that and even right now positioning Honeywell in the AI and robotic side is super beneficial. So I think Honeywell's getting a lot of value just in the previous life that you had, which I think is really great.

Andrey Shtylenko (10:09):

Yeah. Possibly. Hopefully.

Joe Toste (10:13):

Hopefully? So there's actually a pretty great book out there that I think all MBA students have to read. I know Amazon's Jeff Bezos requires his senior managers to read this book called The Goal: A Process of Ongoing Improvement. It's been out for 35 plus years, which influenced industry leaders to move towards continuous improvement. Quick question: Have you read the book?

Andrey Shtylenko (10:34):

I've heard a lot about this book. It's on my reading list. Yeah.

Joe Toste (10:38):

Okay. On the reading list. Love it. So there's a concept called business process management, BPM, which aims at cost reduction and efficiency improvements. AI is taking the place in both the operational and strategic BPM. When it comes to the next generation of warehouse robotics at Honeywell, can you talk about how Honeywell is transforming the way work is done with robotics and industrial IoT in warehouses today?

Andrey Shtylenko (11:03):

This actually, in the recent months, has been area of my focus, so I love talking about that. The warehouses and distribution centers, they're going through a massive transformation these days, of course, due to the extreme demands that e-commerce is putting on the existing logistics network. Existing logistics staff, many times, is very conventional, very outdated, very human labor demanding. I've been through many warehouses and DCs in the last several years. We've been to everywhere from the DCs that were literally a built up a couple of years ago, and the ones that were built 30 years. Those a very different. And the labor shortage, there is a huge problem. So regardless of whether you might have seen some examples of contemporary warehouses that are fully automated, also known as "dark warehouses," the majority of conventional warehouses, they still heavily rely on manual labor.

Andrey Shtylenko (12:01):

So these are not vacation homes. This is a very hard laborer's job. So worker's safety is a big factor there. We see the greatest opportunity for robotics in jobs where automation can reduce or eliminate the risk of injury, repetitive motion discomfort. And especially now with COVID-19, we've got a brand of totally new use cases, from compliance related to personal protective equipment, there's infections, and use cases like that. So if a warehouse, if a distribution center, does not have an automation staff, they're likely just going to be out of business soon. But the good thing is warehouse and DC space is a great fit for a robotics. So these are historically pretty controlled environments, and under those conditions, advanced robotics is thriving. And so, our goal is to develop solutions that would free and support workers from sort of uncomfortable and dangerous tasks, and allow them to take on more sort of satisfying and fulfilling roles.

Andrey Shtylenko (13:05):

And what's been happening in recent years with the technology advancing: new sensors, actuators, [inaudible 00:13:12] robotic platforms, and new embedded compute platforms, and of course with additional machine learning and simulation. That all paves the way for a dramatic transformation of existing solutions. So we are playing to our own strengths here, having the main experience understanding of industry demands, huge install base, established network of vendors and partners. We are collaborating a lot with, for example, Carnegie Mellon's National Robotics Engineering Center, MIT Media Lab, on areas with a heavy research scope. And, of course, our investments and startups like Soft Robotics. And then, of course, bringing on our cloud platform, Honeywell Forge, and our own robotics platform [inaudible 00:13:55] simulation and machine learning capabilities. So that's brings up a very interesting use case. And putting all this together, we were able to move away from the conventional robotics applications, like the ones that were used only in [inaudible 00:14:08] and very constrained environments like manufacturing, [inaudible 00:14:12] less constrained environments like warehouse.

Andrey Shtylenko (14:13):

So our focus is building a robotic solution that can learn and continuously adapt to new scenarios, new types of products, new types of trailers, new setups, new types of lighting; and, of course, operate more safely, operate not only replacing people in the warehouse, because this is still not feasible, but operate together with them. And we do this by our own cloud platform, robotics platform. We collect data from the field from operation, and then use this data to tune our perception models, to tune our planning and reasoning models. Over time, that brings productivity gains. And I can probably bring here some parallels with Tesla, for example. You might know that in April this year, Tesla has released the software update for Model S, and with this new update, the acceleration one from 2.5 seconds to 2.3 seconds without any mechanical updates, only with software.

Andrey Shtylenko (15:14):

So this is where we're trying to go with a similar story. And so, we have this robotic trailer unloader. That's probably a good example to [inaudible 00:15:24] here. The trailer unloading floor for distribution centers is one of the most demanding, injury-prone jobs, and there's a lot of heavy lifting, seasonal extreme heat and cold, high turnover. And so we have the solution, which is pretty much a robot that drives into the trailer and start picking and sweeping packages and pulls them into the conveyor. This robot can perform in any weather, it can reduce injuries, increase the capacity, and with adding machine learning, adding simulation, we are able to localize better within the trailer. We are able to identify different types of packages. And when we put a simulation here, we are able to, with time, build better prediction models, reason-finding models, efficiency, we have the efficiency gains out of more continuous work of the solution at the customer side.

Joe Toste (16:23):

Oh, that's so great. So I'm actually super curious what you said a little bit earlier. What are your thoughts on warehouses and COVID-19 and social distancing? Where does Honeywell see that opportunity right now?

Andrey Shtylenko (16:36):

So we are definitely exploring the new opportunities here. I won't be able to talk [inaudible 00:16:42] to lengths of what are the products that we are working on, but that's definitely one of the new constraints that's been a huge impact to the industry. But again, we see a couple of use cases here. There's, of course, the robots for disinfecting, there's, of course, monitoring for compliance for the personal protective equipment, and, of course, the monitoring compliance for the distance. And this only relates to robotics, right? But it also relates to the algorithms that could be implemented into sensors and surveillance cameras and surveillance systems.

Joe Toste (17:24):

That's great. One transition into AI and robotics, I know it's super fascinating to you, I'm just curious: What makes it so interesting to you in today's marketplace?

Andrey Shtylenko (17:36):

Probably I should say that AI and robotics is pretty different fields. So it would be probably mistake to just talk about both of them from this one perspective, but there's actually a huge stereotype that, so far robotics is all AI, but this is a huge overestimation. What we see in our workflow is probably 10-15% of [inaudible 00:18:00] robotics, it's probably any sort of advanced algorithm. So I originally transitioned from the software background, and I started many, many years ago building desktop applications, websites, web applications, mobile applications. When looking to what I've done before that, I would never probably go back to software, for a couple of reasons.

Andrey Shtylenko (18:25):

First of all, it brings a new paradigm of engineering. So instead of just coding and engineering a rule-based system, you build the skeleton, you build sort of a platform, and then you teach that platform to operate within a specific context. So in the case that I brought before, our robotic unloader, we have the hardware, we have the skeleton, and now we have to infuse and then train on the environmental data for it to become more efficient. Almost like you're raising the child, and I know that you can relate, as well. There's definitely some parallels how you teach your child and how you teach a machine. It's really an interesting experience.

Andrey Shtylenko (19:04):

And then, probably the second reason is: I love the fact that you can just build a successful data-driven solution without closely interfacing with the customer. So you can build software, you can build apps, you can build websites by the specification, but building the data-driven solution, specifically solutions that are expected to continuously learn from the experience without going and learning everything about the user context, is almost impossible. It's like trying to teach a child English by speaking Russian. You have to know everything about the incoming data to build a good perception [inaudible 00:19:41] appraising and planning [inaudible 00:19:43]. And, of course, building AI solutions, data-driven solutions, building robotic solutions, is much more challenging than building conventional software, compared to machine learning or robotics. It's much more challenging to build [inaudible 00:19:57] probabilistic solution rather than sort of deterministic solutions.

Joe Toste (20:01):

Yeah, that's really good. And I like the part where you're talking about we both have young children. And I was actually talking to.. We have an AI practice at our company, and we were talking about this, and how babies in the beginning, there's that decision stage, right? And they're not quite there yet. They don't recognize a ball, for example. They can't tangibly make sense of it in those early months. And then, as they start to move along, they can kind of make that decision that, "Oh, this is a ball," and then, they can kind of make it actionable like, "This ball bounces," and then, "You can kick this ball." Which is really funny because my son, right now, the only ball I give him is a basketball, and so he's got a little basketball hoop and now he's... It's so crazy to see the decision, and then him taking action on it. Not that I want to compare a toddler necessarily to AI and that, but it's just a really great, simple example about how you can kind of start to train the kid, train the model.

Andrey Shtylenko (21:03):

Right, right, right. Yeah. And this is where, in our machine learning field, we have the term "imbalanced dataset." So if you are teaching your kid on some specific objects or samples more than others, and then you see how they learn more from the samples that you are exposing to them. So there's definitely a lot more parallels to that.

Joe Toste (21:25):

Yeah. And that, too, they're learning pretty... I think every day, it's something new. And right now, it's climbing boxes for my son in the kitchen. So I'm just curious: What's your vision for the connected worker and Industry 4.0 with AI and robotics as we move forward in this kind of post-COVID 19 and 2020 and beyond.

Andrey Shtylenko (21:49):

So if you look at what generally what happens within technology [inaudible 00:21:53]: advances in [inaudible 00:21:55], advances in wireless protocols, advances with sensors, software, ML algorithms, human-computer interfaces, this all leads to a lot more sensor-ization, right? So you have lot more sensors, and this leads, of course, to a lot more data collection and processing. It's either processing in the Cloud or it's processing right there on the edge, on the device, or on-prem. First of all, it leads a lot to real-time monitoring, real-time control, and real-time decision support, and the response time is going down. It all leads to improve efficiency and safety for those use cases. It leads to a lot more remote work, and in many times, for example, you don't really have to be on-site to be doing something.

Andrey Shtylenko (22:46):

For example, one of our solutions, we have this augmented reality solution, which is called TechSight, which is pretty much the augmented reality glasses that somebody on-site, when something is happening, something disruptive to the process, whether it's at a warehouse or distribution centers, and when you need to react fast, so anybody from existing site engineers can put on the augmented glass, and then they can translate the feed with the support center, where they can be directed from the engineering side to perform some sort of updates or adjustments [inaudible 00:23:27] fixes. So you don't really have to be on-site to do something. We are moving from more a reactive approach, reactive solution, when accident already happened to more like anticipatory prevention.

Andrey Shtylenko (23:44):

Let me give you an example. So we have another solution for preventing occupational hearing loss, which is one of the most common work-related injuries. So conventional approach would be regularly measuring the level of the noise at the site, and then ensure that the workers that are present there are not exposed to the noise above the required level. So we have a solution which has pretty much a massive noise canceling headset that monitors the level of noise for the worker. And so we can know how much the worker is really exposed to the noise, and we can prevent them from the hearing loss. And then, just by this, you can do things like noise heat maps, where you can preemptively notify the safety manager that the worker is approaching the required [inaudible 00:24:38] and the worker have to take a day or take two days off.

Joe Toste (24:42):

Yeah, I think at the rate... And especially with COVID right now, the rate of change and just digital transformation is just accelerating so much that I think, even the next couple of years I'm thinking about this, and there's going to be some pretty, pretty awesome stuff. I mean, obviously both on the consumer side and on the industrial manufacturing side. There's going to be a lot of change. And this time's really exciting. Great space to be in. Lastly, before we hit, what I call, the 60-second TechTables segment, and everyone tells me, "Joe, I have more than one problem," but if there was one problem you are seeking to solve as the global AI leader at Honeywell and the robotics lab, what would that be?

Andrey Shtylenko (25:22):

That's exactly [inaudible 00:25:25]. That's the first thing that came to my mind is I have the long list of things to [crosstalk 00:25:31] to be very focused. Well, my main objective is, literally I'm counting that in terms of dollars, so my goal is enabling an extra billion dollars of revenues for Honeywell by engineering data-driven functionality in the coming years. Now, this breaks down into many things like building capabilities, bringing in new talent, research partners, solution vendors, ideating and filtering potential ideas. But most importantly, it's probably the continuing to work for the shift of making organization accept data-driven mindset. And it all starts with ensuring that each new solution is enabled for connectivity, does have data and analytics strategy, does have proper computer and data collection functionality so we have the processing infrastructure and [inaudible 00:26:24]. So that data-driven mindset that the whole organization does have to accept is probably the major topic on my agenda.

Joe Toste (26:34):

I love it. Okay. So 62nd TechTables number one: What do you know now that you wish you had known at the beginning? It can either be personal, business, with your startup, with Honeywell, blended mix.

Andrey Shtylenko (26:47):

This is something that I probably touched on already, but allow for more focus on business opportunity and the problem at hand rather than the technology. And in my early years, I was so excited about engineering. I was so excited about the coding. I was too distracted by how I would do something, by what programming language I would use or what kind of library I would use, instead of what problem I'm really solving. And that's because I failed several times within my first startups, because I was making product decisions based on the technology instead of making them based on what the customer wants. And now, when we have new ideas coming in, the first question I'm asking, "Are we really solving the problem here? If yes, can we solve this problem without machine learning? Can we solve this problem using [inaudible 00:27:36]-based approach? Can we solve this problem using logistic regression?" And this goes in contrast to the approach of, "Let's do just deep learning model and see how that performs." So first thing you need to think about is, "What is the business opportunity here?"

Joe Toste (27:52):

Wow. Yeah, that's actually really, really great. I never actually hit 60 seconds on the TechTables. It always goes way over, because everyone always gives me a really great answer. But I love what you said about the customer problem. I just wanted to highlight that with a neon sign. "What problem are you actually solving" is so, so important. Back in the day, everyone wanted to build an app, because that's what they thought they needed, and that no one asked, "Hey, what problem are we actually solving?" It might not actually be an app. It might not actually be anything related to that. Just as a quick example, but I really like focusing on the customer problem and does ML, does AI, does it actually solve that? So just as we're wrapping this up, we're going to transition a little bit. A little bit of an easier question for you. What's your favorite Netflix show right now? This is question number two on the 60 seconds that we [crosstalk 00:28:47].

Andrey Shtylenko (28:46):

I wish I had all the time. I wish I had 72 hours of my day. But I don't watch really much of TV, especially in the last years. But the last show that I watched with my wife was Devs by Alex Garland. Did you watch that?

Joe Toste (29:03):

I haven't. I have not.

Andrey Shtylenko (29:04):

It was fun, because I'll tell you why. The first episode there is, without spoilers, this quantum computing company, and then there's this Russian guy who was doing the AI. And we were watching through the first episode, and my wife says, "Hey, don't you think your use case is the same as this person in this episode?" Because I'm Russian, and Honeywell is in quantum computer [inaudible 00:29:30]. And this guy is killed in the first episode, so that was fun.

Joe Toste (29:38):

Your wife, she sounds pretty funny. It sounds like she has some jokes. That's hilarious. Yeah, I don't normally get to watch any personal TV, but I am often dragged in with... My wife and I have a ten-year-old, and so we'll watch Star Wars with her, and I love Star Wars. And then, every once in a while, my wife will love to watch a couple shows. But yeah, typically, I'm so... We're going to talk about this as we roll into this last question, or at the end about personal productivity. But yeah, it's a struggle for me to pull away, because I'm always loving, but always want to be as productive as possible. The last question: What's your top three personal development books right now that you would recommend to the audience?

Andrey Shtylenko (30:22):

Within the last several months, the ones that I really like to... So definitely How to Fail at Almost Everything and Still Win Big, and this is by Dilbert cartoonist Scott Adams. The one I liked was The Road Less Stupid by Keith Cunningham, I believe. And, of course, Master Your Code by Darren Gold. I try to not generalize and sort of suggest personal development books, because personal development books, they are like a pill. Whatever problem you have, you have to pick up the book related to the problem, right? But those are three that I really like, so [inaudible 00:31:03] more generic ones.

Joe Toste (31:04):

That's great. Yeah. There's a lot of personal development books out there. I think over time... When I was younger, I was reading everything, and now you start to kind of realize, yeah, exactly, someone will give you a pill for... It's like going to the doctor. They'll just give you some meds left and right. So that's it. That's it, officially, for the podcast. Where can people come out and find you? I know you hang out on LinkedIn a lot. Is that the best place?

Andrey Shtylenko (31:30):

LinkedIn is the best place. I totally left the other platforms like Instagram and Facebook, but I'm very active on LinkedIn. This has been a great practice of trying to compile my ideas, my thoughts, and suppose sort of more like a personal development exercise [inaudible 00:31:49] than anything else.

Joe Toste (31:51):

I love that. For the audience: we're actually talking about a personal productivity webinar, me and Andrey doing one together. So we're just going to tease that right now, out in the future. I'm super passionate about that. I know you're super passionate about that. And so, I'm just really looking forward to that happening in the future. And yeah.

Andrey Shtylenko (32:12):

[inaudible 00:32:11].

Joe Toste (32:12):

Yeah. It's going to be really fun. There's a great quote from... It's a leadership podcast I listen to. The guy's name is Craig Groeschel, And he says, "When the leader gets better, everyone gets better." And I'm constantly thinking about that. I really, really like that. Well, that's it for today on the TechTables podcast. Thank you for coming on, Andrey. Really appreciate.

Joe Toste (32:33):

If you're interested in seeing what Nagarro, a high-end technology solutions company to some of the world's leading organizations, can do for your business, you can email Joe at: joe.toste@nagarro.com, J-O-E, dot, T-O-S-T-E, @nagarro.com, or message Joe on LinkedIn. For all information on Nagarro, check out nagarro.com. That's N-A-G-A-R-R-O.com. You've been listening to the TechTables podcast. To make sure you never miss an episode, subscribe to the show in your favorite podcast player. If you have an iPhone, we'd love for you to open the Apple Podcast app and leave a review. Thank you so much for listening. To catch more TechTables episodes, you can go to techtablespodcast.com. And to learn more about our sponsor, please visit nagarro.com. That's N-A-G-A-R-R-O.com. And, of course, you can find Joe Toste, your podcast host, on LinkedIn, Twitter, and Instagram. That's Joe Toste: T-O-S-T-E. Thanks for listening.

Joe Toste
Joe Toste
Host of TechTables Podcast

I'm passionate about investing in communities locally and internationally across several organizations (Young Life, Compassion, and DP Basketball 🏀 (high school basketball coach). This passion intersects across technology and the public sector too.