This is Part Two of a two-part course. Part One is a prerequisite.

A Catholic Introduction to Artificial Intelligence, Part One.

Course Outline:

Part Two: Applied AI & Building with AI

Class 1: Introduction to LLMs Students revisit the inner workings of large language models with a deeper dive into tokenization, context windows, and prompt-response pipelines. They explore how LLMs generate text by predicting the next word and how context length limits coherence. Comparisons between proprietary and open-source models prepare students for self-hosting.

Class 2: Using Open-Source LLMs with Ollama Students are introduced to Ollama, a tool that simplifies running LLMs locally. They learn how to install Ollama, load models like Mistral or Llama 3, and query them from the command line. Students gain insight into the hardware requirements, performance trade-offs, and offline benefits of local LLMs.

Class 3: Open WebUI Frontend – Basics In this class, students install and launch Open WebUI as a user-friendly interface to interact with local models. They explore its chat layout, prompt history, and simple configuration settings. This class emphasizes building user-facing tools with privacy and customization in mind.

Class 4: System Prompts, Roles, and Memory Students explore how LLM behavior can be shaped using system prompts and defined roles. They experiment with prompt engineering to control tone, persona, and output length. The concept of memory (temporary vs. persistent context) is introduced and tested.

Class 5: Tools & Functions with LLMs This session focuses on expanding what LLMs can do using tools and function-calling. Students learn how Open WebUI enables models to access external tools like calculators, file loaders, and custom plugins. They create simple tools and test calling them from within the chat.

Class 6: Prompt Engineering Students develop advanced prompting skills. They learn how to craft instructions with few-shot examples, control outputs with formatting techniques, and guide reasoning using step-by-step prompts. Side-by-side comparisons help students evaluate effectiveness.

Class 7: AI Automation with Make.com In this class, students use Make.com to create workflows that integrate AI models with common tools like email, forms, or Google Docs. They explore scenarios like generating meeting summaries, automating feedback forms, or batch-creating documents with LLM input.

Class 8: AI Agents – What Are They? Students are introduced to the concept of autonomous agents—LLMs that act on goals using tools and memory. They learn about agent architecture (planner, memory, executor), and evaluate tools like LangChain, AutoVPN, and Agen tops. Use cases include research assistants and data scrapers.

Class 9: Build a Simple Agent Students work in teams to design and deploy a basic agent using Ollama + Open WebUI or API-based functions. Agents might be tasked with fetching a file, summarizing content, or interacting with a web form. Emphasis is on prompt design, chaining, and reliability.

Class 10: Group Project Work Session This class is devoted to building and refining student projects. Students work individually or in small teams on an applied AI solution such as a chatbot, productivity tool, or educational assistant—using the tools and models they’ve learned.

Class 11: Project Troubleshooting Day Students focus on debugging prompt issues, interface errors, or tool integration problems. The instructor and peers offer support. Students document challenges and solutions as part of their learning process.

Class 12: Final Presentations I Half of the class present their final projects. Each student or team shares their problem statement, how AI was used, what tools were integrated, and a short live or recorded demo. Peer feedback is encouraged.

Class 13: Final Presentations II & Course ReflectionThe rest of the class presents, followed by a course-wide reflection. Students discuss what surprised them, what changed their minds, and how they plan to use AI responsibly going forward.

•  No textbook; all materials provided with the course at no extra cost.

  This course uses the following AI-based tools:

  ·         NotebookLM is an AI research and study tool developed by Google that helps users understand and work with their own documents. Instead of answering questions from the general internet, it analyzes sources you upload—such as PDFs, Google Docs, or notes—and generates summaries, explanations, study guides, and even audio-style discussions based only on those materials. This makes it especially useful for research, learning, and organizing information because the AI stays grounded in the sources you provide. A Google account is required.

  ·         Jupyter Notebook is an open-source interactive computing environment where users can write code, display results, add explanations, and create visualizations all in one document. It is widely used in data science, machine learning, and AI because it allows programmers to run code step-by-step while documenting their work. This makes it especially useful for learning, experimentation, and sharing reproducible research. Jupyter Notebooks can be created and executed within a browser with no need to install any additional software. No account registration required.

  ·         ChatGPT (https://chatgpt.com/): an AI assistant that explains ideas, answers questions, researches information, and helps with writing, learning, planning, and problem-solving across many topics (from Chat GPT). No account registration required.

  ·         Google Gemini (https://gemini.google.com/app): I am Gemini 3 Flash, an adaptive AI collaborator. I’m here to help you create text, images, music, and video with clarity and wit (from Gemini). A Google account is required.

  ·         Claude (https://claude.ai/login): an AI assistant made by Anthropic. I can answer questions, write, analyze, code, and help with a wide range of tasks (from Claude). Account registration is required.

  ·         Google Teachable Machine:  a web tool from Google that lets users train simple machine-learning models—like image, sound, or pose recognition—without coding. A Google account is required.

  ·         (Optional) Visual Studio Code (VSC): VSC is an integrated development environment (IDE) that aids in the writing, debugging, and execution of a large range of computer programs.  It also can handle Jupyter Notebooks.  Students with computer programming experience and who are familiar with VSC will find using VCS for Jupyter Notebooks quite easy. VSC software is free but it must be installed on the student’s computer or laptop.

  ·         (Optional) Google Colab: (short for Google Colaboratory) is a free cloud-based environment from Google that lets users write and run Python code in a web browser. It works like an online notebook where you can mix code, text explanations, and visualizations—making it popular for data science, machine learning, and AI experiments. Because it runs in the cloud and can provide access to GPUs, students can build and test AI models without installing software on their own computers (from ChatGPT). A Google account is required.

 Students should plan to spend about one hour per class on homework.

Assessment Overview

·         Participation & Discussions – 20%

·         Short essays – 30%

·         Quizzes – 10%

·         Final Project – 40%

For all 13 classes: $237 if you register on or before November 15; $257 if you register after Nov. 15. (Registration closes one week before the first day of class. After that date, registrations are not guaranteed. There is a $25 surcharge for late enrollments after the course is closed.)