My name is Nandakumar Ramachandran Pejari. In this session, we'll dive into how AI and IOT converge to drive real time optimization in multi cloud environments. I'll discuss key technologies use cases that empower organizations to manage IOT cloud infrastructure effectively at a lower cost. In this section, we'll talk about the industry wide adoption of multi cloud infrastructure as well as application deployments And the challenges associated with it. Now, most of us work at organizations or have dealt with organizations. that have started adopting multi cloud as a strategy going forward. Now, this is basically because organizations believe in flexibility. They would want, they would want to avoid vendor lock in and also optimize performance based on applications which are best built for certain clouds. Now, typically customers, big or small, are adopting and placing applications across different cloud providers such as AWS, Azure, GCP, Now, again, the reason for this is to these that strategy offers flexibility and the choice to customers so that they can put applications and infrastructure on clouds of their own choice. However, it comes at with its own challenges. in managing these workloads with the security associated with it and the cost efficiencies associated. Now, do hyperscalers provide a solution to this? Yes, they do. for example, AWS offers something called as AWS Control Tower. Azure offers Azure Arc. Google Anthos is one of the other platforms that Google provides. But they come with its own complexities because these tools are meant for their own specific ecosystem. And thus, there's a role for AI to fulfill over here. On this ecosystem, interconnected services allow tailored strategies. For instance, AWS IoT Core enables edge to cloud IoT data transfer, while Azure IoT Hub provides extensive IoT device management. Now, using Kubernetes and service meshes like Istio ensures seamless integration and orchestration across these platforms. Now, Kubernetes applications tend to be extremely light because they don't carry the burden of having an operating system and thus offer the flexibility and the mobility of placement across various cloud providers. Now, there are other services like AWS Outposts and Azure Stack, which offer a hybrid cloud experience and the flexibility to deploy clouds on prem as well as on prem. On in a hybrid cloud scenario that offer the flexibility. However, again, these come with its own limitations as well. for example, for on prem or for edge use cases, these are not purpose built. They are pretty much seen as a bolt on. Now let's see how AI can be used to solve multi cloud IOT related optimization. Now, IOT generates a continuous stream of real time data. Now protocols like. MQTT, COAP handle a lot of low latency edge to cloud communication for latency sensitive data edge computing platforms like AWS Greengrass or Azure IoT Edge preprocessed data before sending it to the cloud for advanced analytics. So while edges are capable of. data generation, the data actually gets processed at the servers located in the cloud. The flexibility of finding the nearest server and sending the data to it to get it processed is key to low latency when it comes to IoT centric applications. Now, not just that, keeping the applications based on Kubernetes allows for data to be sent to servers that could actually move across multiple cloud providers based on cost and performance requirements. Now, AI drives optimization by analyzing IoT data for predictive maintenance, workload orchestration, and anomaly detection. Now, services like AWS SageMaker. Azure Machine Learning GCP Vertex AI enable building and deploying these models across clouds. AI based orchestration tools like Kubernetes or, for example, like I mentioned earlier, which is pretty much a larger framework of orchestration ensures real time scaling based on runbook automation triggered by a specific alert. These triggers could be based on data volume at edges. data processing requirements at central location and completely based on cost to spin up more Kubernetes pods or even for that matter like, spin down some of the Kubernetes pods when the time comes to optimize workload on when there is a leaner workload in this case. Now the rise of hybrid AI models combine real time inference at the edge with cloud based scalability. For example, smart cities. leverage IoT protocols like LoRaWAN and AI platforms like Azure Digital Twins. to optimize traffic and infrastructure. Now, Kubernetes further facilitates seamless deployment in multi cloud setups with at least, with the least amount of manual overhead, to put it per se. as I mentioned earlier, Kubernetes combined with AI and IoT can simplify a lot of this workload and application orchestration for most of these use cases that we see today across the industry. That's it. Let's talk about agentic AI. Agentic AI is a term we all are hearing a lot of a lot these days, right? So let's see how we can leverage agentic AI to solve some of these Challenges. Now, agentic AI autonomously makes decisions based on real time IoT data. For example, AWS Greengrass combined with, AI models on SageMaker allows devices to, to act locally while syncing with the cloud for broader insights. That's how it is supposed to be designed. That's how most frameworks operate as well. Now, this flow that you see on this slide. This flow shows the interplay of IoT data collection, edge computing, and cloud analytics. Now, services like GCP IoT Core collect real time data, while edge platforms like Azure IoT Edge handle the local inferencing. Now, agentic AI ensures decisions are made autonomously at the edge or in the cloud as needed. Now, as an example, one of the large manufacturing customers that I've been working with, used vibration sensors and MQTT protocol to detect machine performance and failure patterns using agentic AI send. And basically they sent proactive alerts resulting in nearly like 80, 84 percent of lesser downtime. Now this resulted in nearly 12 percent cost savings as well as 8 percent of additional production. Now, if you put it into perspective, if you put these numbers into perspective of a large production setup, this is huge, actually. Some other real world examples include predictive maintenance using OCP UA or, for, Also be UA for industrial IoT and real time supply chain optimization using MQTT messaging as well. Now for smart grid optimization, DDS enables deterministic communication between IoT sensors and cloud AI systems. Now here are some, additional case studies highlighting AI driven solutions, in industries like logistics, healthcare and energy. For instance, AWS IoT Core, Azure IoT Hub securely share patient data in multi cloud setups, while Google Cloud AI improves delivery routes with real time traffic inputs from IoT sensors. In conclusion, I would like to say that AI is critical to achieving real time optimization in IoT related multi cloud deployments. By embracing hybrid cloud models and leveraging tools like Kubernetes, protocols like MQTT and AI services, organizations can drive innovation, reduce costs, and gain competitive advantage. Thank you so much for your time.