Live online Generative AI classes for kids

Class 1: Introduction to Generative AI, overview of Large Language Models (LLMs), how modern AI systems generate text, and the roadmap for the course.
Class activity: AI opportunity map – Students list five real-world situations where Generative AI could be useful and explain how an LLM might help solve them.

Class 2: Running your first LLM locally using Ollama and open-source models.
Class activity: Local LLM experiment – Run an open-source model with Ollama and ask it a few questions to observe how responses are generated.

Class 3: Spanish Tutor demo using open-source models and understanding how LLM applications work.
Class activity: AI language tutor – Use an LLM to translate and correct five sentences to see how the model behaves like a tutor.

Class 4: Setting up the LLM development environment with Visual Studio Code and UV.
Class activity: Development setup – Install Visual Studio Code and open the course project folder to prepare the coding workspace.

Class 5: Installing Git and preparing the development environment for AI projects.
Class activity: Git repository practice – Clone the project repository and explore the files used in the course.

Class 6: Installing UV and creating Python environments for LLM development.
Class activity: Python environment setup – Create a Python environment and install the required packages for running LLM code.

Class 7: Setting up OpenAI API keys and configuring environment variables.
Class activity: API configuration – Add the OpenAI API key to environment variables and test if the program can read it.

Class 8: Installing useful extensions in Visual Studio Code and setting up Jupyter Notebook.
Class activity: Notebook setup – Create a Jupyter notebook and run a simple Python cell to verify the environment.

Class 9: Running the first OpenAI API call and understanding system prompts vs user prompts.
Class activity: First AI prompt – Write a Python script that sends a prompt to OpenAI and prints the generated response.

Class 10: Understanding the OpenAI Chat Completions API and prompt structure.
Class activity: Prompt variation test – Send three differently written prompts and observe how the responses change.

Class 11: Building a simple website summarizer using the OpenAI Chat Completions API.
Class activity: AI summarizer – Provide a long article and generate a short summary using an LLM.

Class 12: Writing an OpenAI API call from scratch using Python.
Class activity: Independent coding challenge – Write the full API request code without using a template.

Class 13: LLM Engineering building blocks, models, prompts, tools, and workflows.
Class activity: AI workflow sketch – Draw a simple pipeline showing how user input flows through an LLM application.

Class 14: Frontier models overview, OpenAI GPT, Claude, Gemini, and Grok.
Class activity: Model comparison – Ask the same question to two frontier models and compare their responses.

Class 15: Open-source LLM ecosystem, LLaMA, Mistral, DeepSeek, and Ollama.
Class activity: Open-source exploration – Research one open-source LLM and explain where it can be used.

Class 16: Chat Completions API, HTTP endpoints vs OpenAI Python client.
Class activity: API method test – Send prompts using both HTTP requests and the Python client.

Class 17: Using the OpenAI Python client with multiple LLM providers.
Class activity: Multi-model script – Write a program that sends prompts to two different models.

Class 18: Running Ollama locally with OpenAI-compatible endpoints.
Class activity: Local vs cloud test – Send the same prompt to a local model and a cloud model and compare responses.

Class 19: Types of LLMs, Base models, Chat models, and Reasoning models.
Class activity: Model type experiment – Run prompts that require reasoning and observe how different models respond.

Class 20: Testing frontier models through web interfaces and research tools.
Class activity: AI evaluation session – Ask three AI systems the same research question and analyze the answers.

Class 21: Introduction to Agentic AI concepts and how modern AI tools perform tasks.
Class activity: Agent idea design – Design a simple AI assistant that could automate a small daily task.

Class 22: Frontier models showdown and building a simple LLM comparison game.
Class activity: Model challenge – Run a creative prompt across multiple models and compare which performs best.

Class 23: Understanding Transformers and the architecture behind GPT and modern LLMs.
Class activity: Transformer sketch – Draw a simplified diagram explaining how transformers process text.

Class 24: From LSTMs to Transformers, attention mechanism and emergent intelligence.
Class activity: Attention example – Highlight key words in a sentence that a transformer might focus on.

Class 25: Model parameters and scaling from millions to trillions in GPT, LLaMA, and DeepSeek.
Class activity: Parameter comparison – Compare parameter counts of several models and discuss their impact.

Class 26: What tokens are and how LLMs process text as tokens.
Class activity: Token counting – Use a tokenizer to count tokens in different sentences.

Class 27: Understanding tokenization and how GPT splits text into tokens.
Class activity: Token experiment – Tokenize different paragraphs and observe token differences.

Class 28: Tokenizing with tiktoken and understanding the illusion of memory in LLMs.
Class activity: Context experiment – Ask follow-up questions to see how models maintain conversational context.

Class 29: Context windows, token limits, and API cost considerations.
Class activity: Token budget exercise – Estimate token usage for processing a long document.

Class 30: Building a sales brochure generator with the OpenAI Chat Completions API.
Class activity: AI marketing generator – Generate a short sales brochure using a product description.

Class 31: JSON prompts and chaining GPT calls together.
Class activity: Prompt chaining – Build a small workflow where one GPT output becomes the input for another call.

Class 32: Business applications of LLMs and building a simple AI tutor prototype.
Class activity: AI tutor mini project – Build a basic AI tutor that answers questions about a specific topic using prompts