🦄 Making great presentations more accessible.
This project aims to enhances multilingual accessibility and discoverability while maintaining the integrity of original content. Detailed transcriptions and keyframes preserve the nuances and technical insights that make each session compelling.
Overview
📖 AWS re:Invent 2025 - Reinventing the Retail Edge with YUM! Brands, AWS & Spectro Cloud (HMC204)
In this video, Kyle Goodwin from Spectro Cloud, Shiv Adhiappan and Ryan Good from Yum! Brands, and Justin Swagler from AWS discuss their collaboration to reinvent retail edge architecture. They explain how Yum! Brands modernized its in-store technology across 40,000+ locations in 155 countries using Amazon EKS hybrid nodes powered by Spectro Cloud, creating a unified Kubernetes management system. The solution enables real-time AI/ML workloads at the edge, reduces IT overhead, and allows level one help desk staff to resolve issues quickly. Key benefits include improved uptime, faster deployment of new features, and enhanced customer and employee experiences. They emphasize the importance of treating edge infrastructure as an extension of the cloud, building modular and future-proof systems, and starting small with proof of concepts before scaling.
; This article is entirely auto-generated while preserving the original presentation content as much as possible. Please note that there may be typos or inaccuracies.
Main Part
Introduction: Three Organizations Unite to Transform Yum! Brands' Digital Journey
Good morning, everybody. How is everybody doing? Good, good. So my name is Kyle Goodwin. I'm the senior vice president of sales and operations for Spectro Cloud. I'm excited to have with me here today Shiv Adhiappan, the vice president of Global Reliability Engineering for Yum! Brands, as well as Ryan Good, director of Reliability and Platform Engineering, and Justin Swagler, AWS Worldwide Head of Physical Retail. So we're actually here today to discuss how we together as three organizations have reinvented retail edge. So we'll kick off with them actually telling us a little bit about themselves. So Shiv, we'll start with you.
Thank you. Hi everyone, I'm Shiv Adhiappan. I lead the global reliability engineering at Yum! Brands. As you all know, Yum! Brands owns Taco Bell, KFC, Pizza Hut, and the Habit Burger, and I've been with the company for the last five years, focusing on the site reliability engineering, incident management, and scalability for our platforms.
Yeah, I'm Ryan Good. I lead our reliability and platform engineering, so a lot of our developer enablement, but also infrastructure, all the fun stuff that actually sells the pizza, the chicken, the tacos that we have.
And lastly, Justin Swagler, part of AWS. I serve as worldwide head of physical retail. So any and everything as we think about technology experiences, operations, and innovation across physical retail, restaurants, and consumer goods, I'm owning from that side from an AWS perspective.
Excellent. Thanks guys. So Yum! Brands runs thousands of locations globally, and tech innovation is happening not just in the cloud, but also at the edge. So today we're actually going to explore how Yum, AWS, and Spectro Cloud are reinventing that in-store architecture to deliver agility, AI, and scalability. So let's start with you, Shiv. What sparked this transformation in the first place at Yum?
Yeah, Yum! Brands is the largest restaurant chain in the world. They operate out of 155 countries. The journey for selling our food product has been transforming into more of digital, and five years ago when I joined Yum, Yum! Brands decided to kind of own its own technology, build their own first-party technology platform to run the restaurant ecosystem end to end. As you do that, you're building your cloud platforms, your digital platforms. In-store becomes such a key part of it because it's a customer journey at one end and there's the restaurant operations and our store operators' experience on the other end. So it basically comes together to form the ecosystem and we've been building the modular product ecosystem over the last five years and we've scaled a lot of our products we call by Bay. It's scaled across like 20,000 plus restaurants already.
Building a Unified Architecture: EKS Hybrid Nodes and the Power of Strategic Partnerships
So how is that, and I think you and Ryan can tag team this a little bit, but how is that increased the speed of your operation, uptime? Has it impacted the uptime agility overall? Can you talk a little bit about that?
Yeah, for sure. So I think especially with the way that we think of platform engineering and we think of building platforms that are self-service, right, one of the key things is you have these operations issues in the restaurants that can now be handled at the level one help desk level. They can see and have visibility, but then also the tooling to take action to restart workloads that might be hung up and restart those POS devices that throw up errors before even having to escalate it to some of those level three or those higher level engineers, so you stop that cycle much faster and so you get them back into business a lot quicker. So yeah, absolutely huge unlock for that.
Yeah, I mean to add to that, operating a restaurant, taking orders means even if Internet goes down, cloud goes down, the last mode for us is running the order taking process and it's important that the in-store technology supports that. It's the backbone of our sales.
Excellent. So from AWS's perspective, Justin, what are some of the key architectural elements of the services that make it possible to run AI/ML workloads?
Yeah, really underpinning that comes down to creating a unified kind of management system as you think about Kubernetes and your applications across any channel, right? And underpinning that is Amazon EKS hybrid nodes which AWS announced last year at re:Invent this time and it's really accelerating the journey for customers like Yum, other retailers that I engage with around the world as they think about creating that unified management experience, right? Yum! Brands, what, 40,000 locations have lots of locations, right, across edge devices and all those applications, right? That's a lot of locations that have to be managed as far as doing modernization, deploying new applications. And with EKS hybrid nodes powered by Spectro Cloud, that gives you kind of a single pane of glass, right?
A control plane that gives you visibility into what you're running, where you're having issues, and how to resolve them quickly. This allows you to modernize in days or weeks rather than months or years, which is what other customers have dealt with in the past. This is fundamental because from a retail and consumer industry standpoint, shoppers engage across all channels: mobile, digital, voice, and physical locations. By having that unified control plane, you can better manage the customer experience from any touchpoint a shopper goes through.
The second key aspect involves a resilient edge and AI applications deployed to those locations. If I do model training on SageMaker or with Bedrock in the cloud, I can deploy lightweight AI models or applications directly at the edge that are custom-fit to the workload and use case I'm trying to solve. By having that hybrid unified architecture, customers like Yum! Brands can bring these applications into their stores.
We talk a lot about being better together across the industry. From a collaboration perspective, this spans Yum! Brands, Spectro Cloud, and AWS. Each partner plays a specific role in shaping the overall solution. We couldn't build everything ourselves. While we own the first-party technology and platform, we had to choose the right partner solution for our architecture and the ecosystem we're building. AWS has been a very strong partner over the years, and what we do is work closely with them to bring our solution roadmap and architecture together. AWS does an excellent job connecting us and doing the matchmaking with appropriate solutions. It's not just one solution; you have all these options you can try out.
We had to start thinking about in-store and edge workloads not as separate entities but as part of the cloud ecosystem. The second you start thinking about them as something different, you start fragmenting your solutions. That's where bringing in a Kubernetes-based solution becomes more critical for us. Having a thought partner in AWS that can bring us to different solutions is huge. We can't see everything and don't have relationships with many vendors, so AWS acts as the matchmakers. With the ecosystem we've been building, the same teams that build for the cloud also need to be able to build for the edge. With that partnership and single pane of glass, the same teams can build workloads and AI across our entire ecosystem.
From an AWS standpoint, partners are fundamental to how we help customers like Yum! Brands by providing the right ISVs and partners like Spectro Cloud to tackle challenging problems underpinned by AWS and Amazon services. We also want to be a strategic partner with you. Beyond just the Spectro Cloud side, we think about bringing in AI capabilities like computer vision, IoT, and agentic AI inside your containers. AWS can build on top of that with you, powered by Spectro Cloud, to create that unlock. It's truly a one plus one equals three proposition long term.
From Foundation to Innovation: Deploying AI and Real-Time Analytics at the Edge
Connecting you with other customers doing similar things is also valuable. When we think about reliability and scaling these solutions to the edge, we can't be everywhere. The example of level ones being able to click buttons and restart pods eventually needs to be automated. Going out and connecting with customers already doing these things, learning from them, understanding best practices, and knowing what models to use—that's where getting in those rooms and making those connections is key. That's actually a good segue to what we see a lot in the industry: aligning enterprise needs with edge and AI innovation. What lessons did you learn in that process from a Yum! Brands perspective?
One of the things we wanted to do is ensure that Edge, as a solution that gets deployed into the restaurant and stays there for five to ten years, is built for modernization and scalability. So one of the key ideas that Ryan, myself, and a couple of our team members discussed early on, before going into the solutioning phase, is that we have to extend the edge into the cloud. Think about a world where your edge servers are also containers—just a node running in the restaurant—and what it would take to scale the workloads from edge into the cloud. That was the key underpinning of futureproofing our technology.
Otherwise, five or ten years later, when all this new hardware comes in that's much more efficient, you just can't truck roll everywhere to recycle all those things. So that was definitely a tricky challenge. We needed to keep the cost low if you do want to extend it, build it modular in a way that works, and make sure that it's future-proof enough to leverage the infinite scale of the cloud that you normally want to use there.
I think to build on that with what you've done here with Spectro Cloud, that's fundamental. That's part of your foundation. If we want to bring in AI and all these cool agentic capabilities, but we don't have the foundation set up correctly, we're setting ourselves up to fail. A lot of our customers are realizing the potential of modernizing, better connecting and unifying their edge and Kubernetes into the cloud so that we can streamline that, reduce IT overhead costs by a significant amount, have full resiliency, low latency, and no downtime.
That foundation then allows us to start doing proof of concepts, starting small and scaling from there, driving further value. It unlocks a lot more workloads that we haven't thought about yet. We have use cases that we're building based on, but now you have the infrastructure in place. The time to market for new functionality and features drastically reduces when you have the infrastructure ready to go. All of the hard work we're doing here is really about a few things, but we'll focus on one: improving the customer experience.
Knowing that Yum! Brands is prioritizing AI and GPU workloads overall, can you give us some concrete examples of how AWS and Spectro Cloud together are actually doing that for Yum! Brands? One of the big things is real-time data analytics. Right now, especially if we don't have this stuff running at the edge, we're having to ship that up to the cloud to do analysis. By the time you get that information down to those store managers or franchisees, it may not be relevant anymore. It could be a week old, and there's nothing they can do with that information.
As we look to the future of being able to analyze that data in real-time locally, store managers can start to take action. Then you start to build agents and automatically take action, which allows those store managers and store employees to focus on what matters: the customer experience. If you think about retail, I remember back when I worked in retail and my manager would be in the office running reports constantly, disappearing every twenty to thirty minutes for an hour and coming back with the latest sales report. What if all of that was just automated and they had the actions they could take? They could spend time with customers, greet them, and really enhance their experience coming into your retail store or restaurant.
Then you have other line employees who can focus on making better quality food and focus on the things that matter to them rather than worrying about operational issues. Those things can really hamper operations. If you take all of that out of their mental load and they just have smooth operations, customers are going to benefit, employees are going to benefit, and everyone's going to benefit. The customer experience is one thing, but the team member experience is equally important to us. That's how you get the best food out to customers as fast as possible.
Our goal was always to make the technology work, like email. Someone said it before: it should just work. How do you make that happen? You build the right infrastructure underneath it, and then you remove the technology from how things work so it's not in your face. You're not having your restaurant team members become their own tech people, trying to unplug and replug wires and things like that. The operational benefit that comes out of that has been huge for us. There are definitely business use cases we can run from an AI and ML workload perspective.
We have our own internal containers and core containers because we could do so much more observability and streamline the connectivity back into the cloud. Rather than having 20 different devices shipping logs to the cloud, we funnel through one node up. In consumer industries like retail, restaurants, and consumer goods, we're providing real-time proactive visibility. If an agent or your containers can tell me about an issue before I, as a restaurant manager or store manager, even know about it, and I can respond to that, that's driving value.
If your cookers fail and you don't know until after it happens, that's bad. By deploying this kind of modern unified infrastructure, we can start to predict issues ahead of time so I can resolve them before they occur, which drives the experience. The operational stability part means that rather than just saying something's wrong, it tells you exactly what is wrong. Then you're jumping into the recovery part as fast as you can, and that's critically important.
Looking Ahead: Modernization Strategy and the Future of Cloud-to-Edge Architecture
Yes, digital commerce is growing, but even younger generations still want the experience inside stores. So you can really focus on that experience, making sure it's seamless, it's connected, and that it's providing a great experience for your shoppers. Since you're on the mic, technology is rapidly evolving, especially as we keep talking about with AI and data-driven retail overall. How do you see cloud to edge architecture adapting and evolving over the next few years to keep retailers like Yum actually ahead of the curve?
AI is going to be a big area that's coming over the next couple of years. We're starting to build that now with customers as we think about physical locations. A lot of the research that we're seeing across the industry shows that 90 percent of retailers are going to be investing in AI algorithms at the edge. They're going to be deploying that to reduce costs in IT and improve workforce efficiency. Even if they save 30 to 40 minutes of their time in the day, they can focus on the experience. Those are the trends that are shaping how customers like Yum need to architect and the types of applications they need to build.
The evolution is going to be starting with that whole real-time infrastructure. I have that infrastructure, I'm now capturing IoT data and all this other data. I'm able to analyze it in real time and respond, whether it is solving an issue, whether it's personalization, retail media, engagement, and so on, to really create a much more seamless experience not just for the shopper but also for your workforce as well. There are two parts to this. One is the small language models that run locally. The other one is the infrastructure for connectivity that allows you to take a big workload, ship it out to the cloud asynchronously, get the response back. You need that edge component to be able to play that role as well.
We're coming up on time, so let's do a quick rapid fire. What's one piece of advice you'd give to other tech leaders trying to modernize and not get left behind? I think understanding the ecosystem and adding those puzzle pieces is going to be such a key part. It's so hard to take all these pieces and try to make them work together. We do an ecosystem diagram where everything just needs to connect and loop back in. Nothing just stays on the site somewhere. That's such an important vision part of it. Building things is now easy. The thinking and strategy part is the hard part now.
One piece of advice would be to lean in to just trying things. You can't be afraid to just try something new and put it out there. The strategy thing is exactly it. Think about all of the things you're trying to solve and look at it as a multifaceted piece, not just a single slice. Think about all of the aspects of that thing that you're trying to solve: customer, employee, business. All of that is part of it. Hopefully, by leaning in on trying some of those things and realizing you're not going to get it right, but if you build it in a way to be modular, you can enhance it later.
One of our Amazon principles is around starting small. Test it out, do a proof of concept, explore. Be curious around what technology like a hybrid approach from an edge and cloud can unlock for you. Start small and then scale from there. But also, modernization like this is a meaty challenge. It can feel overwhelming, like you really need to completely change what you're doing. That's why it's important to start small, try that out, and slowly start to slice away at bringing those capabilities with a strategic view long term. Starting small helps set the foundation for allowing you to really accelerate overall modernization.
What you've heard today is really the power of three organizations coming together for one mission, and that mission was to help Yum deliver consistent, reliable, and innovative experiences to their customers. AWS provides that global scalable foundation, Spectro Cloud delivers that operating layer from a Kubernetes perspective, and Yum brings the vision, really driving with that strategy that pushes all to reinvent. If we can make infrastructure invisible, we can accelerate transformation.
; This article is entirely auto-generated using Amazon Bedrock.
Top comments (0)