Shivay Lamba
Shivay Lamba
Shivay Lamba is a software developer specializing in DevOps, Machine Learning and Full Stack Development. He is an Open Source Enthusiast and has been part of various programs like Google Code In and Google Summer of Code as a Mentor and has also been a MLH Fellow. He is actively involved in community work as well. He is a TensorflowJS SIG member, Mentor in OpenMined and CNCF Service Mesh Community, SODA Foundation and has given talks at various conferences like Github Satellite, Voice Global, Fossasia Tech Summit, TensorflowJS Show & Tell.
Giving Superpowers to Your React Apps with Machine Learning
React Summit 2023React Summit 2023
12 min
Giving Superpowers to Your React Apps with Machine Learning
Have you ever questioned whether Javascript is a viable alternative to Python or R for creating machine learning models? After all, a 2019 survey by Stack Overflow found that Javascript is the language that developers use the most. Given that machine learning models like neural networks require a lot of computational power and that javascript was not intended for high-speed computing, this approach seems unreasonable. But hold on, this not entirely true, as javascript libraries like Onnx.js and Tensorflow.js are here to save the day! I'll be going into further detail on how to create intuitive and innovative machine learning applications with React in this talk.
Using MediaPipe to Create Cross Platform Machine Learning Applications with React
React Day Berlin 2022React Day Berlin 2022
20 min
Using MediaPipe to Create Cross Platform Machine Learning Applications with React
This talk gives an introduction about MediaPipe which is an open source Machine Learning Solutions that allows running machine learning models on low powered devices and helps integrate the models with mobile applications. It gives these creative professionals a lot of dynamic tools and utilizes Machine learning in a really easy way to create powerful and intuitive applications without having much / no knowledge of machine learning beforehand. So we can see how MediaPipe can be integrated with React. Giving easy access to include machine learning use cases to build web applications with React.

Machine Learning based Unit Tesing in JavaScript
TestJS Summit 2022TestJS Summit 2022
22 min
Machine Learning based Unit Tesing in JavaScript
The talk covers the current scenario of writing test cases in JavaScript and the problems associated with the time and resources spent by companies to write the test cases and the lack of automation in this area.
Then the talk will cover how AI and machine learning is being leveraged by tools such as Github Copilot and Ponicode to autogenerate test cases thus simplifying the software testing process.

Server-Side Rendering Using WebAssembly
React Advanced Conference 2022React Advanced Conference 2022
12 min
Server-Side Rendering Using WebAssembly
This talk shares how to achieve Server-side rendering using WebAssembly and WASMEdge which is a WebAssembly Runtime. The talk also covers the benefits of using WebAssembly to achieve Server Side Rendering. The talk will also cover a demo on how to launch a React application using the WasmEdge runtime.

High Performance Node.js Powered by Rust and WebAssembly
Node Congress 2022Node Congress 2022
8 min
High Performance Node.js Powered by Rust and WebAssembly
In the post Moore’s Law era, due to limitations of the hardware, we need to squeeze more performance from the existing hardware. That means that the native code provides the best performance. However, the prevalence of native code on the server-side presents challenges to application safety and manageability. The rise and advent of Rust and WebAssembly offers new ways for developers to write high performance yet safe Node.js applications.
In this talk, I will cover the basics of Rust and WebAssembly, as well showcase how to go about their integration with Node.js. You will learn how and when to design a hybrid web application. How can you code the high performance functions in Rust in a Web Assembly virtual machine and finally how to tie everything together in a Node.js JavaScript application.

Predictive Testing in JavaScript with Machine Learning
TestJS Summit 2021TestJS Summit 2021
18 min
Predictive Testing in JavaScript with Machine Learning
This talk will cover how we can apply machine learning to software testing including in Javascript, to help reduce the number of tests that need to be run.
We can use predictive machine learning model to make an informed decision to rule out tests that are extremely unlikely to uncover an issue. This opens up a new, more efficient way of selecting tests.

Using MediaPipe to Create Cross Platform Machine Learning Applications with React
React Advanced Conference 2021React Advanced Conference 2021
21 min
Using MediaPipe to Create Cross Platform Machine Learning Applications with React
This talk gives an introduction about MediaPipe which is an open source Machine Learning Solutions that allows running machine learning models on low-powered devices and helps integrate the models with mobile applications. It gives these creative professionals a lot of dynamic tools and utilizes Machine learning in a really easy way to create powerful and intuitive applications without having much / no knowledge of machine learning beforehand. So we can see how MediaPipe can be integrated with React. Giving easy access to include machine learning use cases to build web applications with React.
Hello everyone. I'm Shivay Lamba. I'm currently a Google Summer board mentor at MediaPipe, and I'm going to be talking at React Advanced. So excited to be speaking at React Advanced on the topic of using MediaPipe to create cross-platform, machine learning applications with React. So a lot of this talk is going to be centering around machine learning, MediaPipe, and how you can integrate basically MediaPipe with React to create really amazing applications. So without wasting any further time, let's get started.
Of course, I mean today machine learning is literally everywhere. You look at any kind of an application, you'll see machine learning being used there, whether it's education, healthcare, fitness, or mining for the sake of it, you'll find the application of machine learning today in each and every industry that is known to human kind. So that makes machine learning so much more important to also be used in web applications as well. And today as more and more web applications are getting into the market, we are seeing a lot more of the machine learning use cases within web applications as well.
[01:28] And let's actually look at a few of these examples that we can see. For example, over here, we can see a face direction happening inside of the Android. Then you can see the hands getting detected in this iPhone XR image. Then you can see the Nest Cam that everyone knows is a security camera. Then you can see some of these web effects where you can see this lady and she has some facial effects happening on her face using the web, or you can also see the Raspberry PI and other such kind of micro-based microchips or such kind of devices that run on the Edge. And what are the things in common in all of these? That's the question.
What's MediaPipe?
[02:03] So the thing that is common in all of these is MediaPipe. So what exactly is MediaPipe? MediaPipe is essentially Google's open source, cross-platform framework that actually helps you to build different kinds of perception pipelines. What that means is that we are able to basically build or use multiple machine learning models and use them in a single end to end pipeline to, let's say, build something. And we'll also look at some of the common use cases very soon. And it has been previously used widely in a lot of the research based products at Google, but now it has been made sort of upstream and now everyone can actually use it since it's an open source project and it can be used to process any kind of an audio, video, image based data and also sensor data.
And it helps primarily with two things. One is the dataset preparation for different kind of pipelines within machine learning and also building basically end to end machine learning pipelines. And some of the features that are included within MediaPipe include end to end acceleration because everything is actually happening on device. Then secondly, is that you just have to actually build it once and different kind of solutions, including Python, JavaScript, Android, iOS, all those can actually be used. So you just have to build it once and you can use it on different types of platforms. That is why we are calling it a cross-platform based framework. And then these are just ready to use solutions. You just have to import them and integrate them into your code and it'll be very easily used. And the best part about it is that it is open sourced. So all the different kind of solutions, all different code bases, you can find on the MediaPipe repository on Google's organization on Github.
[03:46] Now looking at some of the most commonly used solutions, some of the most well-known solutions include the Selfie Segmentation solution that basically is also actually being used in Google Meet where you can see the different kind of backgrounds that you can actually apply, the blurring effect. So what it does is that it uses segmentation mask to only detect the humans in the scene and it is able to extract only the information needed for the humans. And then we have Face Mesh that basically has more than 400 plus facial landmarks that you can put, and you can make a lot of different, interesting applications using this. For example, let's say AR filters or makeup, right? Then we have Hair Segmentation that allows you to segment out the hair. Then we have standard computer vision based algorithms, like object detection and tracking that you can do to detect specific objects. Then we have facial detection. We also have hand tracking that can track your hands and you can probably use it for things like being able to use hand based gestures to control, let's say, your web applications. Then we have the entire human post detection and tracking that you could probably use to create some kind of a fitness application or a dance application that can actually track you. Then we have the Holistic tracking that actually tracks your entire body, right? And it tracks your face, your hands, your entire pose, right? So it's combination of basically the human pose, hand tracking, and the Face Mesh. Then we have some more advanced object detection, like the three detection that can help you to detect bigger objects like chair, shoes, right, table. And then we have a lot more other kind of solutions that you can actually go ahead and look at. And these are all end to end solutions that you can directly just implement. That is why MediaPipe solutions are so popular.
And just to look at some of the real world examples where it's being actually used. So we just spoke about them. The Face Mesh is a solution that you can see over here taking place on the AR lipstick, Tryon that is there on YouTube. Then we have the AR based movie filter that can be used directly in YouTube. Then we have some basically Google Lens services that you can see like augmented reality taking place. Then you can also see it being used not only in these augmented reality or these kind of things, but also more other kind of inferences like the Google Lens Translation. That also does use the MediaPipe pipelines in its backend. And you can see Augmented Faces that again is based on the Face Mesh.
Live Perception Example
[06:05] So let's look at a very quick live perception example of how basically it actually takes place. For this, what we are going to be doing is we're going be looking at the hand tracking, right? So essentially what we want to do is that we take an image or a video of your hand, and we are able to put these landmarks. What are landmarks? Basically landmarks are these dots that you see and you can superimpose them on your hand and they sort of denote all the different... Like you could say the different edges of your hand and you're going to be superimposing them. So this is what the example is going to be looking like.
So how would that simple perception pipeline look like? So essentially, first you'll take your video input, then basically you'll be able to reduce the frame. Basically you'll be getting the frames from your video and you'll be breaking down that entire frame into a size that is usable by the tensors because internally, MediaPipe uses a TF Lite that is TensorFlow Lite. So you're working with tensors. These are high-dimensional, numerical arrays that basically contain your entire information about the machine learning. So basically you'll be doing a geometric transformation of your frame into a size that is being used by the tensors. So these images will get transformed into the mathematical format of the tensors, and then you'll run the machine learning inference on those. Basically, you'll be doing some high level decoding of the tensors and basically that will result in the creation of the landmarks. And then you'll be rendering that landmark on top of the image and you'll get that output. So essentially what will happen is that if you have your hand and you import those landmarks on top of it, you'll finally get this result that you see is basically the hand tracking.
[07:45] So this way we can build such kind of pipelines and basically what's happening behind the scenes or under the hood is that we have the concept of graphs and calculators. So if you are aware of the graph data structure of how the graph has edges and vertices, similarly, a MediaPipe graph also works in a similar manner that whenever you're creating any kind of a perception pipeline or a MediaPipe pipeline, right? So basically it's consisting of, you could say, the graph in the nodes and the edges. And the nodes specifically denotes the calculator. Now, essentially the calculators are the C++ configuration files that essentially store what exact kind of transformation or what is the main brain. You could think of the calculator as the main brain behind that solution that you're implementing.
And then essentially these are the nodes and the data which actually comes into the node and it's processed and comes out of the node, all of those connections via the edges are sort of what is representing the entire MediaPipe graph. So including the edges, and then what is the input put at the calculator and what is the output. So input is what is coming into the calculator. And once the calculations have been done, once the transformations have been done, what's coming outside. So essentially that is how you can think of the entire perception pipeline of using different kinds of calculators together to form, let's say, one particular solution. And all of that will be represented through this MediaPipe graph. So that's essentially what is the backend or what's going on behind any kind of this backend structure of a MediaPipe solution.
[09:23] Now you can also look at some of the docs to get to know more about calculators, graph by going into, or you can also actually visualize different types of perception pipelines, let's say. The one that we used was actually a very simple one where we were just using it to detect the landmarks on your hand. But if you have much more complex pipelines, you can actually go ahead and use this visit that and look at some of the pipelines that are there to offer on this particular set.
How can you integrate MediaPipe with React?
[09:56] And now coming to the essential part, what this talk is really all about, and that is how can you integrate MediaPipe with React, right? So there are a lot of NPM Modules that are shared by the MediaPipe Google team. And some of these include basically Face Mesh, face detection, hands, basically the hand tracking, Holistic that is having the Face Mesh, hand, and your pose. Then Objectron that is the 3D object detection. And then we have the Pose, right? And we have Selfie Segmentation that we had covered, is basically how the Zoom or the Google Meet background sort of works.
So for all of these, you'll find the relevant in NPM packages and you can refer to this particular slide and you can also look at the real world examples that have been provided by the MediaPipe team. These are available on CodePen. So you can refer to any of these to look at how basically that has been implemented. But what we are going to be doing is we are going to be specifically implementing this directly in React.
[10:00] So here is a brief example of how it's supposed to be working. So in the first piece of code that you can see at the top, where we have basically integrated or we have imported React, we have also imported the webcam because the input stream, right, that we are going to be putting up is with the help of the webcam that we are going to be using. So we have just integrated the webcam. Then we have integrated one of the solutions over here as an example and that is the Selfie Segmentation solution that you can see where we have imported from the MediaPipe Selfie Segmentation in NPM Module. And we have also integrated the MediaPipe camera utils. So this is to basically fetch the details from the camera, right? We do also have some other utils that help you to actually create the landmarks, which we'll discover in a bit. But after that, you can see basically the code where we have used the actual MediaPipe Selfie Segmentation.
And again, the best part about this is that you are not supposed to be writing 100, 200, 300 lines of machine learning code. And that's the benefit of using MediaPipe solutions, that everything is packed into this code. And we are doing such kind of important and such kind of essential machine learning based things like object detection, object tracking that usually run into 200, 300 lines of code. And you can simply just put it in less than 20 to 30 lines of code. Over here, we have just simply created our function for the Selfie Segmentation, where we are using the webcam as a reference. And we are using, on top of that, Canvas as a reference because the webcam is sort of the base, right? You get your frames from the webcam and then you are using the Canvas element on top of it to render the landmarks, right? And over here, you can see that we are just implementing the CDN to get the MediaPipe Selfie Segmentation solution. And then we are rendering the solutions. We are rendering the results on top of whatever is being detected.
Demo Time
[12:42] But yeah, I mean, so far, it's all been sort of discussion. We'll move quickly into the code with the demonstration thing. So if everyone is excited, I'm more than happy to now share the demonstration for this.
So let me go back to my VS code. So over here, basically I have implemented a very simple React application. It's a simple created React application, right, that you can use very simply. You can find it on the Facebook documentation. Over here, in my main app.js code, what I've done is that I have integrated four different types of examples, right? So these four examples include the Hands, the Face Mesh, the Holistic, and the Selfie Segmentation solution. So I'll be quickly showing you the demos of all of these, but I am just going to be quickly demonstrating how easy it is to be able to integrate such kind of a MediaPipe solution or machine learning solution, right?
[13:35] So within my function in my app, I have, for now, commented out all the other solutions. The first one that I'll probably demonstrate is a Face Mesh. So I have imported all the components for each one of these, and currently I'm just rendering or I'm returning the Face Mesh component. So if I go very quickly to my Face Mesh component, over here, I see we walk through the code. You can see that I've integrated, imported React. I've imported some of the landmarks. Now, basically, whenever we are talking about, let's say, the face, right? We want our right eye, left eye, eyebrow, our lips, our nose, and all. So these are basically all the ones that we have imported specifically from the Face Mesh.
And then we have created our function, right? We have created basically the MP Face Mesh that will be used to render the Face Mesh on top of our webcam. So over here, we have just to return, again, the CDN and we are using the facemesh.onResults to render the result that will be seen. So we start off by basically getting the camera. So we use the new Camera object to get the reference of the webcam and using that, what we do is that we wait. We have basically created the acing function since the machine learning model itself that will be loaded can take some amount of time to load. That is why we have just used an acing function to wait for the landmarks to actually load. And that is why we send to the Face Mesh the webcam reference that is basically your current frame that you're using. So once your camera loads, the frame starts coming in. We send that frame to our Face Mesh where it can actually render the landmarks.
[15:17] So basically on the const onResults function, what we have done is that we have taken our video input. Then on top of it, we are rendering the canvas element, right, using the canvasCtx. And what we are doing is that we are going to be now rendering the facial landmarks on top of the actual frame that we are seeing. So that is what you see over here, using basically the draw connectors. This is a util that has been provided by MediaPipe, and we are using that. So very quickly, what we're doing is that we are rendering all the different landmarks and we are finally returning our webcam and also the canvas that is going to be put on top of our webcam. And then finally, we are exporting this React component, right, in our app.js.
So very quickly, jumping into the demonstration, right? So I'll open up in incognito mode and I'll go very quickly to where it'll open up the webcam and we should be able to actually see... As you can see, hi, everyone. That's the second me. And very soon, I should be able to see the Face Mesh actually land up on top of the demo. As you can see, that's the Face Mesh. Boom. Great. And as you can see that as I move around, I open my mouth, I close my eye. You can see that how all the facial landmarks are happening. So I can close this and very quickly change to, let's say, one another demonstration. Let's say, if we can use this time, the Selfie Segmentation and in this, what I'll do is I'll basically comment out my Face Mesh, and I'll just comment this out.
[16:55] And I'll comment out this Selfie Segmentation and I'll save it. And til then it's loading, you can see the Selfie Segmentation. Over here again, we have used it. And what we are doing over here is that we are going to be providing a custom background, right? So that background that we are going be providing is going to be defined over here when basically we are using the canvas.fillstyle. So it'll basically not colorize your human body, but it'll colorize all the other things. So that is why we're using the fillstyle to basically, let's say, add a virtual background.
So if I go again, if I quickly go back to my incognito mode, I can go and look at and very soon I should be able to, if it... Hope everything works fine. So as you can see, this is my camera and my frame is coming in and very soon it should load the Selfie Segmentation model.
[17:49] And let's just wait for it. So as you can see, boom, blue background. So essentially again, how it's working is that it is taking in your body and it is segmenting out your human from the frame. And it's basically coloring the rest of the entire background with this blue color, because it is able to segment out your human body and just color the rest of the other things.
So similarly, you can try out various kinds of solutions that are there. Of course, for my demonstration, I showed you the Face Mesh and also I showed you the Selfie Segmentation, but you can also try out all the other ones that are shared inside of the NPM modules that are provided by the React code. So that essentially is what I wanted to particularly show with respect to the demonstration. Again, it's super quick what I have just shared with everyone, right? Even with the Selfie Segmentation code, the actual logic behind the actual Selfie Segmentation that we are writing is literally not more than from line number 10 till line number, probably till, I guess 36. So within 70 to 80 lines of code, you are really creating such kind of wonderful applications, then you can just think about what kind of amazing applications that you could probably think of and you could actually create with help of MediaPipe. And these are just two of the examples that I have shown you.
[19:13] So I mean, the sky is the limit and the best part is that you don't really need to know what's happening behind the scenes. You don't need to know what kind of computer vision or machine learning is happening. You just have to integrate it. And that is why today we are seeing MediaPipe being used in so many live examples, in productionized environments, by companies, by startups. So it's really the future of web, right? And it's just easily integrable with React that is the future of the web as well.
So with that, that brings an end to my presentation. I hope you liked it. You can connect with me on my twitter @howdevelop or on my Github on And if you have any queries with regards to MediaPipe or being able to integrate MediaPipe with React, I'll be more than happy to sort of help you out. And I hope that everyone has a great React Advanced. I really loved being a part of it. And hopefully, next year, whenever it takes place, I'll meet everyone in the real world. So thank you so much. With that, I sign off. That's it. Thank you so much.

Lightning Talks, Day 2, Node Congress
Node Congress 2021Node Congress 2021
13 min
Lightning Talks, Day 2, Node Congress
Hey, welcome back everyone. We are going to go into a nice Q
&A time with our lightning talk speakers. But unfortunately, I forgot to mention this before, but Avital and Jonathan are not able to join us. As you can see, they're from Israel. So that means it's now, I think, like 2 or 3 a.m. for them. So they decided that it was not a good time for them to join. But we do have Shivaay and Alejandro with us. So would you join me on stage, please? Hi, everyone. Hey, thanks for joining us. So great for this thanks talks. English is not my first language, so forgive me. I would like to remind the audience that they can ask questions. Alejandro and Shivaay, we were just in the backstage discussing some things about TensorFlow. You were having a nice discussion. So maybe you would like to share this with the audience? Sure, absolutely. So I was just describing my overall experience and how I got started with TensorFlow. So as any kind of a newbie machine learning sort of interest, I got interested in machine learning very quite early in my journey in my computer science engineering. And I basically started off with a few basic algorithms. But of course, like as you go deeper into things like neural networks, there you get introduced to things like TensorFlow. And what I found was that TensorFlow is a really powerful tool when it comes to a lot of different neural network and image classification, object detection based algorithms. And with that kind of an experience, I got the opportunity to be working with the TensorFlow organization in Google coding as a mentor. And that sort of led me to working a lot with TensorFlow.js. And one of my projects that I made in TensorFlow.js actually got featured on the official TensorFlow channel on YouTube. And that just had like a great impact in my life. And with that, I just wanted to spread a lot more about TensorFlow.js because it is a very wonderful tool for those people who are coming in from these JavaScript background, but they also want to include things like machine learning their projects. Because normally, if you want to get like machine learning based projects into your web application, you would have to learn Python, deploy your machine learning models separately on a separate server, and then get that done. But with TensorFlow.js, it's very easy to just have a few lines of code, and you have a completely off the grill running machine learning based web application at your hands. So that is the sort of the inspiration that led to me working on TensorFlow.js. Cool. And can you share also a little bit about your developer story? Like how long have you been developing? We've discussed this in the start of the program today. What was your answer? When did you start professionally programming? So I started professionally programming in my first year of my college. So I graduated last year from my undergrad degree. So it's been almost three, three and a half years, ever since I've been working in quite a lot of different internships, full time roles as well. So I mean, and with TensorFlow.js, it's been almost more than an year right now at this point of time that I've been using TensorFlow.js as my primary tool. Cool. And what about you, Alain? So my first experience was with a microchip, PIC16F, something like that, with assembly language. And it was during high school. It was a very hard introduction to programming, but you will also get to do a lot of things with hardware, which was very interesting. Then later on that made me change into careers and go study computer science. Yeah. You even switched the whole studies because you liked that so much. Yeah, yeah, definitely. I always thought I would pursue electrical engineering. And yeah, after having a rough first experience with programming, I changed into computer science. Yeah. That sounds familiar. I actually was in school to be an electrician and I really liked it, working with my hands. But then I had to do an internship and it was around February in the Netherlands. So it was this time, yeah. And it was really cold and I had to work in construction and like put down cables and stuff in houses that were not finished yet. So without windows and doors and anything. So yeah. And then I was like, no, this is too cold for me. I'm not going to do this job. And then I became a developer. Yeah. Don't get me wrong. I still like all the topics on robotics, electrical engineering. I'm so passionate about that. But yeah, programming turned out to be a wiser choice in my case. Yeah. Yeah, I can imagine. We have a question from Juan Gonzalez to Shivay. Do you know about examples of cool and useful things that have done with TensorFlow? Sure. Yeah. So as we know that machine learning has a wide variety of applications in all different fields of life. So you will see TensorFlow being used in a lot of different applications in the field of healthcare. So I myself made like a personal AI based physiotherapist assistant that can actually give you guidance on if you're making certain poses or like whatever body exercises that you're doing, if you're doing it correctly or not. Then we are using TensorFlow for things like detecting cancer in skin diseases, right? To be able to detect whether it's cancer. Then using to detect different types of malaria. And then again, TensorFlow was actually used to detect the coronavirus pandemic because a lot of different x rays of the test were taken in and there were like TensorFlow was used over there to find out, okay, whether this particular person can have the coronavirus because a lot of the coronavirus patients had developed certain kinds of issues in their lungs, which could be detected with these x rays using TensorFlow. So it has a wide variety of applications and that is why it really is just, you know, you might be interested in any kind of a field, whether it's like linguistics, robotics, natural language, right? It could be education and it can be used anywhere. That's the magic of TensorFlow. Yeah. Yeah. Finding patterns, basically. Yeah. That's actually, I consulted a company that is using AI. I don't know, except the AI team is a different team and because of the pandemic and working from home, I don't interact with them a lot. I don't know if they're using TensorFlow, but we are detecting early stage lung cancer also with, well, AI. So it could be with TensorFlow basically. And yeah, that's really cool. Like this, I didn't even know that that was possible. So I've been working there for three months now and every day that I feel really proud to be doing these kinds of things. Next question is from Kerry Cool Dude. How can Node improve the TensorFlow backend in terms of performance? Absolutely. So see, one of the biggest reasons why we would want to use TensorFlow.js with Node is that we always, we know that Node has the just-in-time compiler, right? So that makes the performance boost of using TensorFlow.js because again, like TensorFlow.js is essentially running machine learning models in JavaScript, right? And machine learning models, they require a certain amount of competition power to sort of train themselves on any kind of an input data that you give to it. And using the just-in-time compiler that dramatically improves the performance because with your server side that you have set up, like if you have like a dedicated server setup, you can also use things like dedicated GPUs and dedicated CPUs for the application that you're running. So even those have bindings with the native Node compilation that takes place and that improves, drastically improves the performance of your Node.js based machine learning applications. So using the built-in just-in-time compiler and then giving the ability for using external servers with these dedicated GPUs that have a direct combination within the Node.js bindings. That enables us to put up very quick performing machine learning applications. Super cool. Next question is from Alexey. Can we run TensorFlow on Raspberry Pi 4 or will it require more powerful hardware? So TensorFlow can be run on like Raspberry Pi. So you can also use it with Node.js because with Node.js, you have like specific bindings for Raspberry Pi running via Node. If someone has configured Raspberry Pi via Node server, they can also run that and you can natively run these machine learning TensorFlow based models directly on Node on these Raspberry Pi machines. And if you have a very light machine, then we have something called the TensorFlow Lite that allows you to run very basic models directly on the system itself, on these very low powered devices. There could be mobile phones, it could be something like a very, like even Raspberry Pi 1, 2 as well. So even those are supported in TF Lite. And you can just, if it doesn't run that fast, you just have to wait longer. Probably yes. But most of these models are optimized for these resources. But keeping in mind that you cannot run like, you know, models that are going to take you hours and days to actually train, we have to be considerate about that. But it runs, allows you to run most of the off the shelf models directly on these platforms. And if you don't have a powerful machine, then you can always rent a virtual machine, of course, that just spin up, run your model, close it, and then will not be that expensive. And like, again, optimizations are always being made to the current models to make them more efficient for low powered devices as well. And that is why TensorFlow.js is a great way to also utilize because it runs off the shelf on your browser or on your backend without a lot of competition required. Really cool stuff. I'm going to make a little plug right now, because we have some interest now in TensorFlow. And for the people that are watching right now and want to learn more about TensorFlow, Git Nation, the organization behind Node Congress, it's actually been organizing MLConf EU, I think around in November. And if you have to Git Nation multi-pass, you can watch all those talks in retrospect. You can just go Google MLConf EU and find a lot more great content from the talks we had back then. So now we have time for one more question, and it's from Communus. I wonder if you can use TensorFlow.js to analyze the Node.js internals. Yes. So basically, if you want to use so again, this is a really good question. But essentially, what we are doing is that we're evaluating the performance of our Node.js applications. And how basically we can use the performance of the Node.js to enhance these machine learning models. So definitely, we have these internal tools that have been dedicatedly built to evaluate the performance of your TensorFlow models directly. So yes, that's definitely possible. Cool. Cool. All right. Well, guys, that's all we have time for right now. So I would like to thank you for joining us and hope to see you again soon. All right. Thank you, Mehdi. Thank you, Shibai. Talk to you soon. Bye-bye. Bye-bye.