 Good afternoon to you all. This is Sunil representing Ericsson Gaia, which is a global AI accelerator unit for us. So as most of you know, we are the telecommunication infrastructure company. So mostly at the back end, whatever you are using the operators, whether it is a Geo or Airtel or Rodaphone. So we are not into the B2C space. But what we see is in the B2B space. And what we are going to talk in the next five minutes to excite you what kind of problems we are talking about in the AI ML space, where Python is a backbone for us. Just like we are talking about we are backbone for the operators. So Python is backbone for us when we're building the AI and ML solutions. So what we are representing here, we are building a team of 300 data scientists globally, when we have presence in Chennai and Bangalore in India, and in Silicon Valley, and even in our headquarters in Sweden. So what is our main charter for us is, how do you accelerate the execution of the strategy using the AI? We are using a lot of, maybe a lot of hype is there on the AI. But when we see the real world applications, what I will walk you through in the interest of time maybe for five minutes, but we do have our booth G8 for more detailed discussions, I encourage you to come there. And we are trying to build a critical mass of the competence, and we are trying to innovate in the space through automation, evaluation, and growth. So we want to change the portfolio of the Ericsson. This is what we call it as the AI brain. So what we are talking about the interesting point here is, as you see, maybe there are different layers, both from the product side and the service side. What we mean by the product side is, as you see, we are providing the infrastructure of radio, IP, core. At each stage, it's not just only at an application level, do a recommendation or do a prediction. But we are talking about the interesting problems to solve at each level, and that to its scale. Even if you take an example of Chennai, we can easily guess we have one lakh particular towers, whatever we see. So from different operators. And at the end of the day, we are talking about terabytes of data with us, which may be structured. That is one way of looking at it. And on the other side, if you see, we do have a lot of towers and cells which you see physically, how do you ensure the quality of service? So whenever you are making call, you are making sure that you are trusting the network, and it is you are able to talk with the agreed quality of service. So to do that, we have continuous measurement of the activities, right at the radio level, and at the core level, and network planning level, now even with the IoT. So whatever we are seeing, the 2G, 3G, and even 5G is coming up, so that is one part of the story. Second thing is how we are managing the services. So one is just from the infrastructure planning and network planning level, where we are talking about a lot of automation and optimization we are trying to do. At the same time, if I go for a specific example, let's say I want to predict, OK, maybe a lot of crowd is here, and then what kind of quality of service we can promise, and what kind of predictive actions we can take when we have the information, right? It may be configuration changes, or it may be having the specific parameters which are maybe root causes for that, all this we are trying to do with the Python being the back end. If I walk you through a specific use case, for instance, we are talking about the structure data on one side, and even the other side we are talking, we have a lot of infrastructure. How do you maintain it? It's not like a human task. These guys need to climb. If you talk about the field engineer, the tens of meters to see if the infrastructure is intact, but why can't we use the drone images? That is maybe a starting point for us. And a lot of visual intelligence applications, even though we are talking about a couple of applications here, but it's not limited to this one. If I take one example of weatherproofing, where we are doing a live demo of object detection using Python, which we built it, if you see there is an exposed part, because it's in the outside, in the environment, so what we are talking here is how do you make sure that, OK, there are these kind of classes, whether it's bending too much or maybe they are exposed to components, which may affect us finally when we are making really a call or maybe our 4G data. So these are the interesting points which we are talking about, solving it in our real world. In the interest of time, we are not going too many details, but it's not like we are consumers of the Python or maybe the open source community, even we are contributing back. We do have active volunteers who are contributing it as I could learn at Python packages. At the same time, you can visit github.com slash Ericsson, we do have more. 80 plus resources, some may be telecom domain, if you are coming from telecom, you can use it, but there are open source, for instance, if I can quote one example, EVA, Ericsson Video Annotator. So that's one of the, that's built on Python, so you can access those and give us the comments and we encourage you to visit our booth for more detailed applications. Thanks for the opportunity.