Podcast #179 Fast Tracking Industry Transformation

TRANSCRIPT

Ken: Good day, and welcome to episode 179 of our Momenta Digital Thread podcast series. Today, I'm pleased to host Steen Graham, CEO, and Co-Founder of Scalers.AI, a new startup providing full-stack technology solutions at scale.

Before Scalers AI, Steen led key IoT and Edge initiatives at Intel, including as general manager of IoT and Edge AI, as general manager of IoT Ecosystems and Channels. Over the past decade, he's driven the adoption of IoT, Edge, and AI solutions across industries and built developer tools, ecosystems, solutions, and industry-specific products to accelerate technology adoption and deliver business outcomes to enterprises. These efforts contributed to hundreds of thousands of enterprise deployments, design, wins with over a billion in revenue growth for Intel, and adoption by hundreds of thousands of developers. Steen, it's a great pleasure to welcome you to our Digital Thread podcast.

[00:01:37]

Steen: Ken, thanks for having me. I am thrilled to be on, and I am disappointed not to have found all this fantastic content you produced earlier. Over the past few weeks, I've been tempted not to binge all that content. I've got too much work to do to start binging your episodes. But I think this is truly a treasure trove of great content and leaders across the industry.

[00:02:03]

Ken: Thank you for the kind comments. And I don't think that anybody has complimented quite as well as saying we're binge-worthy. Thank you for that. It's right up there with Netflix's new episodes of their television series. But it's also good to cap this off at 179 with you. I've been equally impressed and surprised that our paths have not run across earlier. Of course, via a friend of a friend recently on the venture side, we learned about the great work you've done and are doing for Scalers AI; it was a must-have conversation to have. With that, you'll know if you've listened to a couple of episodes, we're Digital Thread, and we talk about an individual's digital thread. In other words, that one or more thematic threads that define their digital industry journey. What would you consider to be your digital thread?

[00:02:50]

Steen: Yes, Ken, thanks for asking. Because the real transformative moment for me was the advent of this term, the Internet of Things. And contextually, where I was in my career in life was about ten years ago; I was running the embedded sales group at Intel Corporation. And at the time, if you wanted a premier position, you would be in the client computing group or the server franchise. And this embedded sales group was just a portfolio of acquisitions like cable modem business, Ethernet capabilities, small cell, networking capability. And this interesting, embedded business, which was selling computer chips into cars, digital signs, medical imaging equipment, human-machine interfaces, and industrial applications. And once the term IoT was introduced, and at the time, the Intel leader in this newly formed group introduced the concept of 50 billion connected devices. My world irrevocably changed because I had a great business growing in double digits in Asia. But now, the hundreds of thousands of processors we sold into this vast range of fragmented markets were no longer material to these 50 billion connected devices. And the expectations rose for me in my organization to embed computing in the fabric of everything we do, transforming industries and enriching lives. And for the last decade, I've been working an extra 25, 30 hours a week, trying to understand how we take these incredible digital processing technologies and build an ecosystem to drive line of business transformation and learn about all these operational technology industries, which has been incredibly fascinating and rewarding. I'd be remiss to say there's been a lot of pain and suffering and trial and error over the last decade.

[00:04:49]

Ken: It clearly has. And it's interesting, once it's a morass of trial and errors and others opportunity landscape. Think of the wild west of the old. It's interesting. I saw you use the term or heard you use the term ecosystem earlier. As I always define a wordle chart for an individual as I look at their bio and their background - ecosystem plays very strong in yours. I know you had a long run at Intel, overseeing many IoT key efforts there. And a lot of them had ecosystem in the title. What is an ecosystem to you? And why was this valuable to Intel's efforts?

[00:05:25]

Steen: Well, I really learned ecosystem, in Taiwan, and I think there's no greater ecosystem. First, in Taiwan- the contributions Taiwan has made to modern manufacturing. But furthermore, computing in general and watching that community building, the value of that ecosystem is greater than the sum of its parts. It makes an incredible flywheel effect. That's how I kind of envisioned ecosystems, and perhaps more pointedly when I first arrived in Asia and had a big customer event of all our partners. I was a little nervous; it seemed like everybody was a competitor, and putting them all in the same room was it going to be okay? And one of the leaders in the industry of embedded computing pulled me aside and said, "Steve, don't worry about it. Your role here is to make the pie bigger for everyone." And I've taken that to heart. And I think good ecosystems make the pie bigger for everyone.

[00:06:29]

Ken: It's not only an interesting perspective, but the fact that you anchor it in Taiwan, as many of our listeners will know, Advantech is one of the limited partners behind our venture capital fund. And I must say, I have learned a lot about the Taiwanese culture and its impact on both embedded computing and AIoT, as we like to call it. Steen and I have shared many stories about our interactions in Taiwan. It's a good reference, if you will, to have. Going back to the Wordle chart again, if there were another word that I see that makes up Steen Graham, it would be Edge. What did Intel consider to be the Edge when you were there? And why was this important?

[00:07:14]

Steen: Well, Ken, I must interject earlier to your point because you brought up Advantech. And I'd like to take the opportunity to highlight my learnings from Advantech, as well, because coincidentally, KC founded Advantech the same year I was born. And very early in my career, when I was in Taiwan, I was presented with the opportunity to see KC and his peers in the entire Advantech team keynote, produce some incredible thought leadership, and build an incredible ecosystem at Advantech. And so just attribution to growing up on that culture that Advantech built on bringing together amazing technologies and ecosystems. And probably, there's no greater contribution than Advantech made to this term Edge as well. You talked a little bit about this concept around Edge and what Intel considers the Edge, and this is to be quite direct, this brings back many painful memories of hours of whiteboarding and dialoguing. And you can imagine a Fortune 50 company that's so interested in getting the definition right. I was intimately involved in creating the definition of the Edge. We just said it's placing the compute closer to the point of data creation or service delivery point. More important to me has always been why Edge computing how it complements the Cloud or existing modalities of computing and, ultimately, how Edge computing can deliver industry transformation in the end.

[00:08:46]

Ken: When you left Intel, which was not long ago, you collectively led about $2 billion in business across about a thousand partners in the IoT and Edge AI ecosystem and channels. The obvious question is, what compelled you to make the jump into the startups as you did last July?

[00:09:07]

Steen: I must say, Intel was so foundational to my career. And it's hard to outline an number of companies that have done more for humanity than Intel, notably with Pat Gelsinger returning who, early in my career, in my early 20s, I had the opportunities to be in rooms with Pat and see what a brilliant, ambitious mind he was. I was thrilled about the prospects of Intel and the future of Intel. We all look back on our lives and think about the what-ifs. And for me, I grew up in an entrepreneurial family. And that was a passion point for me. I just had a son at that timeframe, and I wondered what he would think of his father if he was a risk-taker, if he took a chance and did what he's passionate about. And then finally, I think I had a few aha moments on these technologies over the past decade. And so, what over the past decade would have taken tens of millions of dollars to do and bring value in the market was taking tens of thousands of dollars if we've done it right. And I got the bug as it pertains to building great as to cloud full-stack software that solves industry-specific problems and the realization that I could do it at a 100x more affordable envelope over time. And in the startup world, you're able to do with much more agility than a large Fortune 50 company with so many great advantages, but the agility and the ability to hit that 100x cost profile just wasn't there. The combination of all those fronts culminated in me resigning from Intel on this incredible company, this incredible opportunity, and founding Scalers.ai.

[00:10:58]

Ken: And that's a great lead-in into scalers AI. You co-founded the company with also, I think, well-known former Intel, Chetan Gadgil, last July, I believe. What's your elevator ride speech?

[00:11:12]

Steen: Yeah, thank you. And I appreciate the shout-out to Chetan Gadgil. He's truly an innovator in the industry. He has a lot of shared experiences over the past decade, both at Intel, where he and I founded initiatives like market-ready solutions and RFP-ready dev kits, and in his prior roles at GE Predix and Cisco. But as you allude to, what we're doing at Scalers.ai, and that elevator ride dialogue is fast-tracking IT innovation into operational technology business outcomes. By complementing new products introduced by that IT industry with the complementary technology and industry knowledge required to deploy the physical world. And that's going to both fast-track product-market fit and OT and enable solutions providers, system integrators, and ISVs to fast track their technology readiness and deploy the latest best technologies so that they can get paid for deploying technologies and servicing technologies faster and sooner with the latest, greatest tech.

[00:12:16]

Ken: My business partner who heads up our venture's investments in the US, Michael Dolbec, came across you through a common connection he had, probably GE Predix and back into Chetan. But Mike paraphrased this as product-market fit as a service. Do you think it's a fair analogy?

[00:12:34]

Steen: This is the journey that I've seen happen so many times. I've worked on many projects, too many to name, whether they are IoT, Edge platforms, developer kits, or all these new products being introduced, whether a message bus, orchestration, or the next greatest computer chip. There are so many new products being built to drive and mine business transformation. And as for those products, research and develop specific industry use cases, the right solution architecture will eventually complement their products with all the code around them, with all the third-party ingredients from the ecosystem, to achieve product-market fit. I've watched good product managers, good alliances, and business development people take two or three years to do that. Ultimately, where I got excited was when at Intel, leading the ecosystem, we were able to take cloud companies and IT companies and give them product-market fit in applications like in-line AI defect detection for marquee-name brand car companies in months, extremely affordably. And so, these are the concepts that Chetan and I tested, and it was just too inspiring to solve that problem that was only increasing. I saw more innovative new technologies being introduced to solve business problems. And more and more people are running through that painful, multi-year, very expensive journey that I wanted to help them out with and fast track. And so, Chetan and I feel like we're well-positioned to help companies solve that problem and fast track their product-market fit.

[00:14:17]

Ken: Well, maybe to put a point on this, I know that you're still somewhat in stealth if I hit your LinkedIn profiles- at the same time, I know as of last week, you were at the ISC West in Las Vegas participating in the Intel booth. What was it that you were promoting in the booth last week?

[00:14:34]

Steen: Yeah, that's a great observation, Ken. Not only were we in the Intel booth, but I was doing a webinar with Aero and Intel, and Microsoft the week prior. I'm passionate about making other companies' technologies shine. Last week at the Intel booth, what we were showcasing was some first-in-the-world technologies. We're the first company outside of Intel to implement some technology that allows you to run Linux and Windows simultaneously and get access to the hardware acceleration through Intel's integrated graphics technology. And why that's so important is Chetan and I and the Scalers team have some great experience with this technology. If you took the same class system as we were running just last week in Las Vegas, the type of AI we did a year ago was basic anomaly detection time-series data. But with access to the integrated graphics through the hypervisor, which is unique, you can run multiple video feeds and algorithms. And what we did in that particular use case, we took these great technologies from Microsoft called EFLOW Edge for Linux on Windows, and that was the great investment that Intel made in that pair virtualization graphics driver. And we built a full smart parking application around it as well. One of the coolest things we did in that use case was a Windows-based video management system. And one of the things is that is interesting is that VMS software is written in Windows. And so right now, you're in this challenging environment where everybody wants to cloud written in Linux, everybody's developing Linux, AI's written on Linux, and everybody wants to add these new modern AI applications on top of Windows. Still, there's no great way to do it. And thanks to the technologies Intel and Microsoft provide now, there's a fantastic way of doing it. And Chetan and I just built the full-stack solution around that. And we were delighted to do that in a few weeks and get it up and running to showcase some incredible partners in that video public safety ecosystem last week in Vegas.

[00:16:45]

Ken: Excellent, and I'm sure, given how long you guys have been around. It does seem that you have a large team behind the two of you at this point; again, I guess another stealth element of how you guys operate. Let's talk a bit about some other early use cases and wins.

[00:17:00]

Steen: Yeah, thanks, Ken, for setting me straight because we have a great team behind Chetan and me, and they were huge contributors to the results there. And there's no way you can, as you alluded to. There's no way you could generate the wins that we have without a great team behind you. Specifically, from a use case perspective, one of our goals at the onset of the company was to have a portfolio of use cases that we feel reflect several of the largest opportunities in Edge computing in the IoT space. And so, the first project that we completed was an in-line AI defect detection solution. Because Ken, you probably saw- maybe you're like me where we thought the predictive maintenance was going to be the first mover turns out replacing visual-auditory inspection techniques.

In some cases, with automotive cars, expensive ultrasound techniques to look at a defect, such as quality defects, were a great use case for in-line AI because you can save so much on scrap and rework. The first solution that we built was in-line AI defect detection for a marquee global Fortune 50 company. We were excited to jumpstart in that space. And since then, one of the other areas that I was excited about- and this is something we showcased in a webinar with Aero, Intel, and Microsoft is we also built a smart port solution. And certainly, one of the challenges we've seen in the world lately is a shipment of containers from China to the port of Los Angeles or Long Beach, where roughly 40% of imports in the United States come through- traditionally would cost $1,500 to $2,000 is now costing $20,000. A huge problem that we're facing right now as far as the supply chain bottlenecks. And we built a great solution with some great underlying technologies from Intel and Microsoft that solves the challenges around bottlenecks within ports, notably how fast the trucks can get loaded and unloaded in a safe way as these ports become more congested as we get more Greenfield crane operators and truck drivers in that space as well. From defect detection to smart ports, I just talked about our showcase at ISC West last week, where the application we're running was a smart parking use case. We're also building medical imaging solutions, a great Edge AI use case. And then retail inventory management solutions, getting the right product, the right place, the right time, it's a $1.6 trillion problem. And then brand management within retail-like Planogram compliance as well. We've had some early wins with customers that have allowed us to build out a nice portfolio of use cases as well, and we're certainly thrilled about that.

[00:19:48]

Ken: To put a point on this, who would you consider your key clients and partners, and how would they engage you? How would they find out more about you?

[00:19:59]

Steen: It's a great question, and so it's really on two fronts. With product companies, we want to work with the leading digital product companies to fast track their product-market fit in these incredible industry-specific outcomes. On one side, a client for us as a product company with a great new Edge orchestration software or time-series database or a great new computer chip that they want to show off, and just fast track their ability to show it off in these incredible new use cases. On one side, that's a goal for us. On the other side, for solution providers, SIs and ISVs, we want to fast-track their ability to deploy the latest technologies, win RfPs, and win new business. Ultimately, both sides of the equation are the right partners for Scalers.ai. And just like we talked about building ecosystems and my early experiences in my career in Taiwan, my expectation for us as a company is to make the pie bigger for everyone. Making the pie bigger for product companies and making the pie bigger for solution providers as well- they are often frustrated by all this new tech and whether it's ready for market and what components need to be stitched up to make a full-stack solution. And we can help with that. And ultimately, that will lead to more sales for the product company. They'll be thrilled as well.

[00:21:22]

Ken: Excellent. And how did they find out more about you? Just hit the website, I imagine.

[00:21:26]

Steen: Yes, hit the website. We're very responsive, and we're looking forward to anybody interested in reaching out.

[00:21:32]

Ken: Alright. Or you can hit Steen or Chetan's LinkedIn profile. But don't be surprised if you see stealth still listed on there.

[00:21:40]

Steen: Ken, you got it.

[00:21:40]

Ken: It's great. The number one thing we advise people to do when they're changing jobs is, regardless of what you're going to do next, if you've got some time off, just put stealth startup because everybody's intrigued. And it's meaningful and meaningless at the same time, yet it's intriguing. Whether you're going to do a startup or not, it's good advice that I'm sure somebody gave you. Let me ask you to put your prognosticator hat on. I noticed, getting Edge, and we're talking about ecosystems, and you mentioned OT earlier. You've been effectively at the forefront of industrial Edge computing for quite a while. This is an area that Momenta is quite interested in regarding both investments and some other work we do. What do you think the next several years hold for us in this industrial Edge computing area?

[00:22:28]

Steen: Well, we're at this incredible point where it's becoming much more common and mature to deploy deep learning at the Edge. All these challenges that we were facing earlier, with the advent of deep learning- and being quite direct, I think most of us have seen deep learning pervasive and cloud applications today, where you've got a huge amount of deep learning, doing memory-based reasoning to serve ads or natural language processing. But obviously, computer vision exceeded the human eye's capability several years ago. But the tools at which to annotate that data, relatively small datasets, deploy those algos, and a full-stack solution create a training loop. They are now much more mature. So, where I think things are interesting is the ecosystem's maturity around deep learning at the Edge. There's no clear platform winner in that space, but the technology is available to implement it. And I think that's what I'm extremely passionate about. And I think having been in this space for the past decade and just connecting the physical world to the internet was certainly transformational. And the data set you got from that was certainly interesting. But deep learning provides us this incredible opportunity to make decisions on this data in near real-time at the Edge. I think that was what was missing earlier at the advent of line of business and operational technology transformation. You and I both know that the half-life of data is shorter at the Edge than in the Cloud. Just being able to make discerning quick decisions with great deep learning technology now and do things like retraining for these practical problems, like getting the camera angle set up correctly, that technology is certainly fast maturing. And that's something I'm certainly thrilled about over the next few years.

[00:24:17]

Ken: Yeah, actually, as we're recording this, the tinyML Foundation is having its annual meeting in the Bay Area. And Mike Dolbec, as I mentioned earlier, is there. But as many of you will know, we recently featured one of their founders, Evgeni, on a podcast. And we're also an investor behind Edge Impulse, which is one of the leading, if you will, MLOps or DevOps platforms for tinyML. We agree fully that this is an interesting space, especially for processor constrained, computing constrained, storage constrained, and bandwidth-constrained devices that are out there which describe many IoT use cases. We're excited about that. Finally, I always like to ask a closing question: where do you find your inspiration beyond KC Liu, who, by the way, I should plug that KC, the chairman of Advantech, participated in podcast number 127. If you want to hear it directly from his mouth, you can listen to that podcast.

[00:25:17]

Steen: KC does a fabulous job. He is truly inspirational. And I did peek ahead and saw that podcast, Ken, and it's now on my to-do list. You probably are not surprised that I find a lot of inspiration through podcasts, and I think it's such a great platform. Certainly, you're contributing with several industries thought leaders and yourself, talking about these critical issues that we're facing and the opportunity to deploy this fantastic technology. I certainly listen to a lot of podcasts. One that I think has been a fun one and a little lighter one throughout the pandemic has been the "All-In" podcast. David Sacks and Chamath Palihapitiya, Jason Calacanis, and David Friedburg do a fantastic job talking about modern events and what's going on in the markets and the tech environment. It's an easy-listening podcast unless Friedburg talks about some deep science topic, which is fascinating, and I must rewind it a bit.

From a book perspective, I think one that stands out, certainly at the formation of a company, is "Principles" by Ray Dalio. Certainly not a recent book, but I think that when you form a company, it's an incredible opportunity to reset and not fall into complacency on how you make decisions. Ray inspired me by his view of presenting everybody in a data-driven way, making inclusive decisions, and seeking out everybody's feedback in an organization to make decisions.

I think where our world sits right now; it's so easy for existing entities to grow and grow and grow in such a way that you want to tell somebody that is in a leadership position what they want to hear. Ray's Principles books did a great job on that. Not to mention the ideas around videotaping decision-making meetings so that the whole organization understands how the decision was made and can now execute it better- building that next generation of leaders that they also get access to.

And then the true personal inspiration for me is my son. I've got a little two-and-a-half-year-old running around. And every day, he inspires me in the mental aptitude of those young kids as they're a sponge for new information. It's just fantastic. I'm trying to keep up with him every day as well.

[00:27:37]

Ken: That's tough to have a child at that age to do these startups because it's a very formative time. And so, in some sense, given our remote working lifestyle, now you both have more time and probably more quality time than you would have been- kind of the corporate gig out. But I could see equally compelling, "Should I go and focus on another 20 to 30 hours per week? Or should I spend the time with my son?" So hopefully, you're able to do both of your passions. Well, Steen, thank you for spending this time and sharing these insights with us today.

[00:28:11]

Steen: Well, Ken, thrilled that you had me on and even more thrilled, frankly, that I discovered the platform and the great content that you and the team have produced. So excited about what you do next year.

[00:28:21]

Ken: Thank you, and I'm thrilled to have you on this episode. So effectively capping off all the others that we've done. So welcome to the Digital Thread club at this point. This has been Steen Graham, CEO, and Co-Founder of Scalers.AI, a new startup providing full-stack technology solutions at scale. Or to attribute to Mike Dolbec, it's a product-market fit as a service. Thank you for listening, and please join us next week for the next episode of our Digital Thread podcast series.

[The End]