Welcome to Jupyter AI. Coding by hand where you write every line of code manually is becoming obsolete. So, coding notebooks have to evolve away from manual coding to using AI to code for you. I'm thrilled to be here with Brian Granger, who is co-founder of Project Jupyter, to introduce Jupyter AI. You can either use it here on the DeepLearning.AI site, or run locally on your own computer. Jupyter notebooks are a workhorse of AI development and data science. And we also use them extensively at DeepLearning.AI. Moving notebooks forward to the AI coding era will be an important step for our field. I'm thrilled to be teaching this with Brian. Thanks, Andrew. When we created the IPython notebook in 2011 and later rebranded it as the Jupyter notebook in 2014, We did so with the mission of creating a community-driven, open-source ecosystem of tools for data science, scientific research, and education. Jupyter notebooks have long been the default prototyping environment for this type of work. But it turns out that most of the AI coding assistants that have emerged in the last couple of years have struggled to function well in a Jupyter notebook environment. And so, we've built Jupyter AI, which is an open source framework that is purpose built to integrate AI into Jupyter notebooks and JupyterLab. AI works well for the kinds of tasks you might already be familiar with from using other AI assistant coding tools, like writing code and answering your questions about code. But it is more than that, and also has capabilities designed specifically for working with coding notebooks. Let me show you. With Jupyter AI, you have a chatbot called Jupyternaut alongside your notebook. So, I can say, Hi Jupyternaut, and get a response. Or maybe more usefully, I can say, write code to implement a coin toss and print heads or tails with 50-50 chance. and it generates the code. Instead of manually copy-pasting it into my notebook, I can just use a single click here to insert the code into my notebook, and look, it runs. Or something I often ask an LLM for help with, if I've forgotten what's the syntax for calling the OpenAI API. I can also ask it to write the sort of static code to get my OpenAI secret API key, and then make a call to the LLM. Or if I'm looking at someone else's notebook, I can drag a cell over here and ask, what does this cell do? or even ask, what does this whole notebook do? and get a quick summary to help me understand the notebook. That all this is integrated with Jupyter makes the workflow of using AI coding with a notebook much easier. You can think of Jupyter AI as an AI collaborator for everything you do in Jupyter, whether that's exploring data in a Jupyter notebook prototyping LLM workflows or even building custom agents that are integrated into JupyterLab's chat experience. In fact, there's much more to Jupyter AI than we'll have time to cover here, but this course will give you a great start. I don't know, Andrew. I'm thinking maybe we need to have a follow-up course after this. There's a lot of exciting things in Jupyter AI that we won't have time to cover today, such as custom agents, tool calling, and agent-to-agent collaboration. No, that sounds great to me. And even in this course, I think you learn a lot. In the first exercise, you learn how to use the chat interface to generate code and ask questions about code. For example, generate a call to the OpenAI API. Then in the second exercise, you use Jupyter AI to build a chat app that can help you find good books using an online books API. And through that, you learn more advanced prompting strategies. Then in the third exercise, you create a workflow to retrieve and analyze stock data and see some of the best practices for how to use Jupyter AI for data analysis. This course is set up a little differently than our other short courses in that it's not arranged as a side-by-side notebook with video for every lesson. Instead, I'll first demonstrate each exercise in a video. and then you get hands-on practice yourself and can build the same application I demoed using the same prompts or modify the prompts to build something customized to your own liking. Many people have worked to create this course. I'd like to thank David Qiu and everyone else that worked on Jupyter AI, and from DeepLearning.AI, Ryan Keenan, David Villarreal, and Hawraa Salami. Let's go on to the next video to get started with Jupyter AI.