5 Unique AI Projects (beginner to intermediate) | Python, LangChain, RAG, OpenAI, ChatGPT, ChatBot

Summary

In this video, Tina, a former data scientist, introduces five AI projects for immediate implementation. The projects range from no-code tools like personal finance chatbots and AI music composers to intermediate and advanced levels involving coding and custom web apps. She discusses using various AI models and APIs, emphasizing prompt engineering and practical applications for personal and professional tasks, including a personal AI travel app and workbooks for books, all aimed at enhancing productivity and leveraging AI's capabilities.

Takeaways

• 😀 The video introduces five AI projects that viewers can start working on immediately, with varying levels of complexity.
• 🛠 Level one of each project focuses on prompt engineering and no-code tools, demonstrating the capabilities of no-code solutions in building simple applications.
• 🔧 Level two involves intermediate steps with some coding required, and introduces the use of Python and APIs for more advanced functionalities.
• 🚀 Level three is for building full-fledged applications, incorporating techniques like Retrieval Augmented Generation (RAG) and frameworks like Flask.
• 📚 The video mentions the use of various AI models, including OpenAI models, Claude, and open-source models like Llama, for different projects.
• 💡 The first project is about creating a personal finance AI chatbot using prompt engineering and no-code tools to help manage and plan finances.
• 🏞 The second project discusses building an AI music composer to generate custom music tracks, which can be useful for background music in live streams or studying.
• 🗺️ The third project is about developing a personal AI travel app to plan customized itineraries, incorporating no-code tools and APIs for data enrichment.
• 🖼️ The fourth project talks about creating an AI that can analyze and generate descriptions for images and videos, with potential applications in learning sign language or understanding social cues.
• 📓 The final project is about generating workbooks for books using AI, which can compile questions and exercises from a book into a PDF for easier use.
• 🌐 The video also mentions tools like Lang chain for implementing RAG, and platforms like Make.com and Bubble for no-code automations and web app building.

Q & A

What are the three levels of AI projects mentioned in the video?

The three levels are: Level one involving prompt engineering and no-code tools, Level two requiring some coding and intermediate skills, and Level three involving building full-on applications and implementing techniques like RAG (Retrieval Augmented Generation).

What is Tina's background and what does she currently do?

Tina used to be a data scientist at Meta. Now she creates content and runs a program called Lonely Octopus, where she teaches people AI and data skills to work on real freelance projects.

What is the purpose of the personal finance AI tool discussed in the video?

The personal finance AI tool is intended to help Tina get her finances back on track by understanding her spending habits and creating a financial plan to achieve her goal of buying a house by the end of 2025.

What is the 5W framework mentioned for creating prompts?

The 5W framework stands for Who, What, When, Where, and Why. It is a method to structure prompts in a way that covers all necessary aspects of the task at hand.

How does the video suggest categorizing expenses for the personal finance AI tool?

The video suggests categorizing expenses into essentials and non-essentials, identifying saving potentials, and optimizing essential spending.

What is RAG (Retrieval Augmented Generation) and why is it used in the video?

RAG is a technique that enhances the accuracy of AI models by allowing them to access a specific data source when interacting with users, ensuring the model references back to the stored information rather than just the original training data.

What is Lang chain and how is it used in the video?

Lang chain is a suite of tools for implementing RAG. It is used in the video to create a custom web app that can handle financial statements and interact based on the user's data.

What is the significance of the HubSpot AI resources mentioned in the video?

The HubSpot AI resources are designed to help users understand and make the most out of chat GPT, including over a 100 prompts to get started with, which can improve productivity and streamline workflows.

What are the additional skills needed for Level two and three of the AI music composer project?

Additional skills for Level two and three include Python, the udio API, the open AI API, text to speech, Dolly for image generation, and a choice of front-end like streamlit.

How does the video suggest generating a playlist of AI music?

The video suggests using AI generation platforms like udio and sunno for Level one, and for Level two, using the udio API with Python to programmatically generate links to the music and compile them into a playlist.

What is the final output of the AI music composer project at Level three?

At Level three, the final output is a full playlist with audio and visual components. The audio is from AI-generated music, and the visual components are generated images related to the mood and lyrics of the songs, created using Dolly and GPT 40.

What inspired the AI-generated workbooks project?

The project was inspired by the desire to have a compiled workbook for exercises and questions from the book 'The Millionaire Fast Lane', which was difficult to manage with scribbles in a physical copy.

What is the process for creating AI-generated workbooks at Level one?

At Level one, the process involves finding a PDF version of the book, feeding it to a large language model, creating a prompt to summarize the book, extracting questions, and compiling them into a downloadable PDF workbook.

What tools are mentioned for creating AI-generated workbooks at Level two and three?

At Level two, the tool mentioned is Bubble, a full stack no-code app builder that connects with different APIs. At Level three, the tool is pdf.co, a low-code and REST API for converting, editing, extracting, and automating PDFs.

What is the potential use of AI-generated workbooks?

AI-generated workbooks can be used for personal use, potentially for compiling exercises and questions from books. However, the video does not provide legal advice on commercial use.

What is the context window of GPT 40 and how does it compare to other models?

The context window of GPT 40 is 128,000 tokens, which is smaller compared to models like CLA that has 200,000 tokens and Google's Gemini which claims to have up to 1 million tokens.