The Function of AI in IoT

Ryan Chacon is joined by SymphonyAI CTO Vijay Raghavendra on this episode of the IoT For All Podcast to debate AI’s position within the IoT trade. Vijay begins by introducing himself and the corporate earlier than speaking concerning the significance of AI in IoT. He then talks extra particularly about AI’s match into enterprise expertise and the right way to begin adopting it. Vijay and Ryan then go extra high-level with conversations round challenges to adoption and the way price impacts the trade’s development earlier than ending the podcast with Vijay letting us know what to search for from SymphonyAI sooner or later.

About Vijay

Vijay Raghavendra is a expertise chief and entrepreneur with in depth expertise main expertise groups in firms starting from startups to Fortune 1. Most just lately, Vijay was the CTO at Acuity Manufacturers, an industrial expertise firm. At Acuity, he was chargeable for all facets of software program expertise technique and supply, together with edge computing and IoT. Earlier than that, Vijay served as SVP of service provider expertise at Walmart, the place he and his groups have been chargeable for all platforms, purposes, and algorithms that drove the expertise for Walmart’s retailers and suppliers and a major a part of the client expertise throughout shops and on-line.

Involved in connecting with Vijay? Attain out on Linkedin!

About SymphonyAI

SymphonyAI is constructing the main enterprise AI firm for digital transformation throughout crucial and resilient development industries, together with retail, client packaged items, monetary providers, manufacturing, media, and IT service administration. SymphonyAI companies have many main enterprises as purchasers in every of those industries. Since its founding in 2017, SymphonyAI has grown quickly, approaching 2,000 proficient leaders, knowledge scientists, and different professionals. SymphonyAI is a SAIGroup firm backed by a $1 billion dedication from a profitable entrepreneur and philanthropist, Dr. Romesh Wadhwani.

Key Questions and Matters from this Episode:

(01:54) Introduction to Vijay and SymphonyAI

(06:23) Function of AI in IoT

(08:57) How AI matches into enterprise expertise

(14:00) How one can begin adopting AI

(18:27) Challenges to adoption

(22:38) How does price have an effect on adoption

(26:42) What to look out for from SymphonyAI?

---

Transcript:

– [Voice Over] You might be listening to the IoT For All Media Community.

– [Ryan] Hiya, everybody, and welcome to a different episode of the, IoT For Al Podcast, the primary publication and useful resource for the Web of Issues. I’m your host, Ryan Chacon. I do ask if you’re watching this on YouTube to please give this video a like, and subscribe to our channel, in the event you haven’t already completed so. And in the event you’re listening to this on a podcast listing, please be at liberty to subscribe, to get the most recent episode as quickly as they’re out. All proper, on as we speak’s episode, we’ve a Vijay Raghavendra, the Chief Know-how Officer at SymphonyAI. They’re an organization that’s constructing a number one EnterpriseAI firm for digital transformation throughout crucial and resilient development industries. Together with retail, client package deal items, and plenty of others. Actually good dialog right here. We speak lots about form of AI, the position of AI in IoT. We speak concerning the applied sciences that they’ve constructed, the applied sciences that they work together with regularly, how individuals can get began with adopting AI into their resolution. Why vertical particular experience is de facto vital, in the case of that sort of integration and bringing AI into an answer, is especially an IoT resolution. And we additionally speak lots concerning the challenges that they see from their facet of issues, because it pertains to bringing options to life. So, all in all nice dialog, quite a lot of worth right here. I feel you’ll take pleasure in it. However, earlier than we get into it Works With, by Silicon Labs has emerged because the go-to developer convention for constructing the talents wanted to create impactful linked units. On September thirteenth by the fifteenth, Silicon Labs is bringing collectively influential, expertise manufacturers, ecosystem companions, and builders for 3 days of technical coaching and workshops, keynotes and professional panels. Works With is reside on-line and free. Register at workswith.silabs.com workswith.silabs.com And with out additional ado, please take pleasure in this episode of the, IoT For All Podcast. Welcome Vijay to the, IoT For All Podcast. Thanks for being right here this week.

– [Vijay] Thanks lots, Ryan. Nice to be right here.

– [Ryan] Completely. Very enthusiastic about this dialog. I wished to kick it off by having you give a fast introduction about your self, to our viewers, in the event you wouldn’t thoughts.

– [Vijay] Nice. I’m Vijay Raghavendra, I’m the CTO at SymphonyAI. I got here to SymphonyAI about seven months in the past because the CTO and previous to Symphony, I used to be the CTO at an industrial expertise firm referred to as Acuity Manufacturers. Acuity, along with being one of many largest industrial tech participant with the lighting and lighting controls additionally has a linked units play for constructing administration, for location administration methods as nicely. And I spent a while working with Acuity to construct these capabilities out. And previous to Acuity, I spent about seven years at Walmart, a number one numerous components of engineering and product at Walmart. And I got here to Walmart as by an acquisition of an organization the place I used to be a co-founder and CTO that I bought my co-founders and I bought to Walmart.

– [Ryan] Incredible. Yeah. Very in depth background expertise. Feels like a reasonably enjoyable journey to get to the place you at the moment are.

– [Vijay] It’s been, yeah.

– [Ryan] So let, let me ask you this. So let’s speak about SymphonyAI actual fast. Inform our viewers somewhat bit concerning the firm, what the main target is, the position you all play in IoT, that form of factor?

– [Vijay] Yeah, so SymphonyAI is an EnterpriseAI firm and our focus is to use AI and machine studying to resolve issues in numerous verticals that we play in. From retail to monetary crime, to industrial to media, IT providers and federal. So, the main target for us is to allow our clients in every of those verticals to actually remodel what their companies and remedy actual issues by the appliance of AI and machine studying. All the work that we do is grounded in our AI platform that we name Eureka. We not solely help the entire capabilities that you’d count on from any of the AI platforms, however we even have some distinctive capabilities round expertise that we’ve constructed particular algorithms, corresponding to topological knowledge evaluation. And particularly for every of those verticals. Considered one of our essential methods by which we add worth, or we carry worth to our clients is thru the deep vertical experience and our pre-trained fashions and capabilities that we’ve throughout these completely different verticals. And particularly coming to the appliance of IoT, relying upon the vertical take industrial, for instance, we work with some very massive manufacturing companies and skim knowledge from plenty of completely different sensors. As you may think, in quite a lot of manufacturing facility to then use the info, to determine, to foretell outcomes corresponding to when a compressor could also be going unhealthy. How do you then do preventive upkeep on these gear to do it, to stop a a lot greater downside from taking place downstream? And with that, we usher in the entire capabilities, not simply with all kinds of various sensors and units that we learn knowledge from, however edge computing, digital twins, and deep studying fashions within the cloud.

– [Ryan] Completely. Yeah, unbelievable overview. Thanks a lot for form of giving us somewhat little bit of context there. I do wanna ask you although, simply from a excessive degree standpoint, after we speak about AI and IoT, they oftentimes at the moment are going extra hand-in-hand than ever earlier than. Inform me concerning the position and the way you view it of an AI firm within the IoT area?

– [Vijay] Yeah, I feel that’s an incredible query. And I feel for the longest time, I actually imagine that folks have at all times believed within the worth of IoT and the potential of IoT. And you’ll see that within the client area. And I feel that choice of IoT within the client area has, clearly, it’s change into mainstream now, however within the enterprise area, we’ve at all times been on the cusp of realizing the worth. And up till just lately, I might argue that we hadn’t absolutely realized the worth, however with the entire adjustments which can be taking place and have occurred, with the price of {hardware} coming down with the density of the compute in edge units. Getting to a degree the place it turns into actually attention-grabbing for us to do quite a lot of processing on the edge with the flexibility now to embed a TPU, for instance, in an edge gadget. So we will run tiny ML fashions on the edge, mix it with working a deep studying fashions or extra subtle fashions within the cloud after which pushing the outcomes again. I actually suppose we’ve the entire underlying capabilities we have to actually carry the ability of AI mixed with IoT to resolve actually attention-grabbing issues, such because the one I discussed just a bit bit earlier. So, I actually imagine that AI turns into the conduit for unlocking the worth from IoT, as a result of with out the flexibility for us to do one thing attention-grabbing and helpful with the entire knowledge that you simply’re getting from these units, it turns into clean. And with the entire adjustments within the {hardware} with cloud, with edge, with the development and AI, I actually suppose it’s an effective way for us to carry the ability of IoT and remedy actual issues.

– [Ryan] One hundred percent agree. So inform me somewhat bit about after we speak about AI, , there’s a number of firms on the market who do AI, say they do AI, there’s a number of options on the market. And oftentimes for the folks that wish to carry AI into their resolution, they don’t at all times know precisely what the distinction is between firms that play within the area that the choices, what they need to be in search of. So inform me, it’s mainly, it’s two questions I’ve. One is to inform me how, what you all do form of differentiates from different gamers available in the market? And the opposite is how does this kind of expertise actually slot in to the way forward for enterprise expertise as an entire?

– [Vijay] Yeah, so, I feel you hit on a extremely essential level with the appliance of AI and enterprises. I feel there are a selection of research that present that 80 plus % of all AI tasks or AI tasks in enterprises which can be attempting to undertake AI fail. And so they fail for plenty of completely different causes. Beginning with actually a lack of expertise of those particular sort of issues there’s attempting to resolve for, for which AI and ML are a extremely good match, not having the precise knowledge or knowledge tradition, and never having the precise perhaps mindset or adjustments within the processes that they should do to actually reap the benefits of this expertise. So, I can maintain going. However the truth is that quite a lot of enterprises battle as we speak with the appliance of this actually attention-grabbing half expertise to resolve issues. So, how will we, or the place will we play and the way will we assist our clients actually get previous this essential problem? As I discussed in my intro, one of many essential locations the place we’ve invested as we’ve constructed out our merchandise, and product providing within the completely different verticals is the deep vertical experience that we’ve constructed over time. So, we’re not a generic AI or ML platform with a couple of fashions and we will throw it over the fence to our clients and say, “Nice, go have at it.” What we’re centered on due to our experience is with each single vertical, we’re tackling the particular set of issues the place AI and ML are an excellent match that. And the place we will go assist remedy these issues in a singular and differentiated method. So take one thing like determining which assortment try to be carrying in a retail retailer and the way a lot of every assortment try to be carrying and the place? That could be a very particular downside that each single retailer has to resolve, large, massive, or small. And what we’ve completed with our deep experience and experiences, we’ve constructed over time these pre-trained fashions that remedy this downside for retail and retailers and CPGs. And with the partnership that we’ve with our clients, we then begin from our pre-train fashions that aren’t simply the fashions and the options, but additionally embed the data of what a service provider at a big retailer does. How do they give thought to what the precise assortment is? What the combo is? We embed that intelligence to then work with our clients, to then optimize these fashions and options to resolve the issues. And we do the identical factor with monetary crime and anti-money laundering, for instance, and so forth. And as in, so doing, we carry some very distinctive IP. I discussed the topological knowledge evaluation. Which is a extremely distinctive and attention-grabbing method to consider knowledge and discover knowledge in an unsupervised studying method. Which takes very massive dimensional knowledge and permits us to suppose, take a look at this knowledge and discover relationships that won’t in any other case be apparent or that ordinary clustering algorithms could not offer you. So, all of those collectively allows us to actually differentiate ourselves from others and give attention to fixing the issues for our clients.

– [Ryan] Yeah, it makes complete sense. One factor you talked about in there that I truly wished to comply with up and ask you about is you have been speaking about form of that vertical, particular experience of the area experience that’s tremendous helpful and form of the differentiator for you. I do know after we speak to, let’s say platform firms within the IoT area, that’s an enormous factor for them to assist separate themselves out is versus simply having this normal platform that may do all of it, or at the very least that’s how they promote it. They discovered worth in buying clients with extra of a focused focus primarily based on area expertise that they’ve for fixing a selected downside or explicit use circumstances inside an trade. So let me ask you when a listener to that is trying to form of be taught extra about getting AI components concerned of their resolution. They’re trying to undertake AI expertise. How ought to they form of go about getting began down that course of and why is it so vital to discover a firm with the vertical particular experience that connects to them to assist simply enhance the chance of success?

– – [Vijay] Yeah. Glorious query, once more. My recommendation to firms that wish to incorporate AI to resolve issues can be for them to be very clear and spend the time to actually perceive that the outcomes they’re attempting to have an effect on and the particular sort of issues they’re attempting to resolve and actually get educated both by partnerships or working with firms, corresponding to ours to essentially perceive the forms of issues for which AI is an efficient match, as a result of it’s not a panacea for each downside that each firm’s gonna have. So, it’s actually vital for them to know that. The second is, and I can’t emphasize this sufficient, AI or any of those machine studying algorithms simply don’t work, or it means nothing in the event you don’t have a tradition of excellent knowledge and a tradition of enthusiastic about knowledge as the important thing enabler for the appliance of AI. If it’s actually rubbish in and rubbish out. So in the event you don’t have good clear knowledge, and if you’re not likely being attentive to that as a core basic a part of the way you construct your merchandise and your methods, frankly, nothing else is gonna work.

– [Ryan] Proper. So, I might say I might actually encourage firms to actually essentially deeply perceive each the issue, but additionally then give attention to the info and making certain that they’re not solely have the precise knowledge, however they’ve a tradition of making certain good clear knowledge as a result of that then turns into an enormous unlock. After which, clearly, specializing in not simply broad enabling platforms, which more and more have gotten a commodity, however working with and partnering with firms that may actually usher in that very particular area experience turns into a key technique to making certain that you simply’re fixing issues that matter as a result of, finally, the constructing blocks for the way you do carry knowledge from numerous sources, clear the info and remodel the info and the way you construct the fashions themselves, are more and more a commodities. The bottom line is gonna be, do you actually perceive the verticals and the domains? So you might be extracting the precise options. You perceive when knowledge or the fashions are drifting. All of these key issues are key issues, I feel are what is going to make it profitable.

– [Ryan] Completely. Yeah, completely. Let me ask you this slight little pivot right here to somewhat completely different space of focus, however while you form of take a look at the evolution of the expertise, not simply in IoT, but additionally in AI and form of how they work collectively, what do you suppose have been a number of the greatest challenges to getting the expertise the place it must be? To extend adoption, to form of, , these expectations or these numbers that we’ve been promised for therefore a few years concerning the development and adoption of those applied sciences, what do you suppose have been the most important perhaps roadblocks or pace bumps which have form of bought in the best way to that form of main as much as the place we at the moment are?

– [Vijay] Yep, I might say, in the event you suppose again to the final decade or so, as I discussed a short time again, I feel the promise whether or not it’s with IoT and even broadly with enterprises has at all times been there, however the adoption, and extra importantly, the success from the adoption of this expertise hasn’t fairly saved up with the expectations. And the explanations are, I feel, once more, a basic, perhaps lack of expertise of how this expertise works. And as I discussed, the give attention to knowledge, and I do know I’m harping on this, like fairly a bit however I’m doing it as a result of it’s so vital. I feel firms, normally, have underestimated the significance of getting an excellent knowledge, however extra importantly, having a tradition of excellent knowledge that’s inherent within the firm. And while you don’t have that, it turns into very tough to understand the worth actually at scale. So, I might say that’s perhaps one of many greatest basic challenges that I’ve seen. The second is I feel, as we’ve developed, particularly within the final 5 to seven years, the capabilities that we now have from whether or not it’s from cloud frameworks to different open supply frameworks, to plenty of Python libraries, to transformer fashions that at the moment are accessible, or has essentially modified the sport. To the purpose you don’t want a PhD in math and stats and laptop science to begin constructing and realizing worth. So, I feel the expertise has additionally developed actually quickly within the final a number of years, that makes it way more attention-grabbing, way more tractable downside to resolve. And let’s face it the final and the most important downside is at all times gonna be for us to seek out good expertise at scale and knowledge science and ML Expertise might be one of many hardest expertise for us to seek out. And that’s the place the flexibility for us to have the ability to leverage quite a lot of these open supply fashions for a citizen knowledge scientist, for instance, to have the ability to use quite a lot of these fashions and options to resolve enterprise issues without having a workforce of information scientists. I feel all of those collectively will assist unlock the worth sooner.

– [Ryan] Yeah. Completely agree. I imply, there’s new applied sciences on daily basis, proper? You realize, BLE Wideband, Extremely-Wideband. Edge computing’s changing into extra highly effective, the cloud. All very large enablers of what we’re form of speaking about. How do you suppose the price component elements into the form of adoption? I imply, clearly price appear to be happening throughout the board for IoT parts tech. Whether or not connectivity, the {hardware}, or the software program, you identify it. How do you suppose that mixed with the evolution and development of the expertise side is taking part in a task and influencing the long run development of what we’re attempting to construct?

– [Vijay] Yeah. I feel price, particularly with IoT within the enterprise is gonna be an enormous issue. Apparently, a number of the work that I did at my earlier firm, and that was simply having a dialog with one of many firms that was utilizing Extremely-Wideband to do look monitoring, for instance. The dimensions for, in the event you take retail.

– [Vijay] The dimensions at which at the very least a big retailer who has a number of areas, the size at which they function, it turns into the price of the {hardware} turns into a really, very materials price, particularly for firms which can be working on very small margins to start with. And I feel that’s the place it must be a mixture of price of the {hardware} has to maintain coming down and we’ve to way more environment friendly concerning the density of the {hardware} or the beacons. And so forth that we’d like in very massive areas. The, as you talked about, edge computing and the density of the {hardware}, the ability of the {hardware} on the sting is changing into an increasing number of amenable for us to do quite a lot of processing on the edge, which then helps us with the egress and ingress cross to the cloud. However, finally, I feel we’re at some extent the place all of those are trending in the precise method. The fee is coming down. The expertise with Extremely-Wideband, BLE, VLC with the NextGen of what’s coming with 5G for the enterprise are all, I feel driving efficiencies and have gotten create enablers for purposes that we will remedy. And finally, if the purposes that we will construct on all of those applied sciences, doesn’t ship sufficient worth to justify the worth then clearly it’s gonna fail because it ought to. However my agency perception is we’re at some extent the place we’ve all of those constructing blocks. The fee despite a number of the challenges, for instance, for a big retailer is at some extent the place it is vitally a lot within the ballpark of being very manageable for a retailer. And the worth that they will get from it, I feel, is justifiable and can solely get higher any more. After which you may apply the identical analogy to numerous verticals as nicely.

– [Ryan] Completely agree. Completely agree. Yeah. It’s all superb factors. I imply, the expansion and the brand new applied sciences popping out is a big half. The fee happening is an enormous enabler. The simply developments throughout the board and all the pieces happening. The extra use circumstances, the extra profitable deployments, the extra area experience that we talked about earlier. Like simply the extra accessible that data and knowledge is, and applied sciences is for individuals, the extra choices they should construct the precise resolution for his or her explicit use case. So, as we wrap up right here, I wished to ask form of trying ahead from the place we at the moment are popping out of SymphonyAI, and form of what you all have happening. What’s the large subsequent step? Like what ought to we be on lookout for being attentive to popping out of your man facet of issues?

– [Vijay] Yeah, so one of many locations that we’re centered on is, as you concentrate on the vertical particular experience and fixing issues in several verticals, particularly with the AI and together with using IoT and different units, the best way we’re enthusiastic about the issues we’re fixing and what’s coming subsequent is de facto the notion of an AI enabled digital employee who’s actually working in live performance in-hand-in hand with the people within the loop, if you’ll. And actually enabling the people who’re working, whether or not it’s a enterprise analyst, a enterprise professional, somebody on meeting line, or a retailer supervisor to do their jobs a lot, way more effectively than they will as we speak with what are perhaps fundamental analytics and even some fundamental AI enabled purposes. In order that’s the place we imagine we’re going to see the actual mainstream, not simply adoption, however a step perform enhance within the worth that this expertise can actually drive in with the appliance. Which brings collectively IoT and the entire ecosystem round it with edge and the cloud with AI and ML, with the vertical experience. Actually bringing all of these collectively in a method that basically augments and helps the people in a basic method. Which makes it in a seamless basic method, I feel is the unlock.

– [Ryan] Couldn’t agree extra. Yeah, it’s very thrilling stuff. What you all have happening over on the firm could be very fascinating from the analysis that I’ve completed. I hope our viewers takes a while to kinda look into what you may have happening. For our viewers on the market who does have questions, could need to comply with up, be taught a bit extra, what’s one of the best ways that they will do this?

– [Vijay] So, our web site symphonyai.com is a superb useful resource. There are hyperlinks there. There’s quite a lot of actually nice content material for the issues we’re fixing in every of the verticals. And there are hyperlinks there for anybody who desires to ping us with any query as nicely. So that will be an incredible place for any of the listeners to ping us.

– [Ryan] Incredible. Effectively, thanks a lot to your time. Actually respect it. We don’t have the chance to speak lots about AI currently. So I actually respect you taking the time to do this. I feel our viewers is getting a ton of worth out of this dialog. So thanks once more, and we hopefully like to have you ever again, different members of your workforce again to proceed this dialogue and speak extra about how AI and IoT actually work collectively to take issues ahead.

– [Vijay] Nice. I respect the chance, Ryan. It was nice to speak to you. Thanks.

– [Ryan] Thanks. All proper, everybody. Thanks once more for watching that episode of, IoT For All Podcast. When you loved the episode, please click on the thumbs up button. Subscribe to our channel and you should definitely hit the bell notifications so that you get the most recent episodes as quickly because it change into accessible. Apart from that, thanks once more for watching and we’ll see you subsequent time.