AI, ML, and Financial Markets Workshop Series

An interactive textbook for the Gies MSBA workshop series on AI, machine learning, and financial markets. Ten chapters in five parts. Every chapter has code you can run in the page, a visual explainer you can poke, a checkpoint quiz, and a Colab notebook for the real thing.

Start here: the seven-minute overview

A walkthrough of what the series covers, how each chapter works, and what to do before the first session. The slides it uses are below the video.

Series overview · 7 min · Download the slides (PowerPoint)

Note: The schedule has changed since this video was recorded. See Schedule and rules for the current dates, times and rooms.

The story

Every chapter is set at Champaign Capital Research, a fictional independent equity research firm that is adopting agents and machine learning one careful step at a time. The firm, its people, its problems, and its plan for the year are on the firm page. The firm’s goal is to automate its investment research workflow with an agent and machine learning, with a person approving every result before it leaves the firm. Every chapter moves it one step closer; read the firm page first.

How this book works

Each chapter follows the same shape so you always know where to look:

1. Objectives — three things you can do by the end.

2. Concepts — the idea, kept short, with a picture you can step through.

3. Try it — a Python cell that runs in your browser. Edit it, break it, run it again. No installs.

4. Checkpoint — one question with instant feedback.

5. Exercise — the thing you actually build, in Colab against a real model on Lumen, the campus LLM service.

The four parts

Part	Chapters	The question it answers
0 · Foundations for AI	0, 0A–0H	How does a large language model work and how is it trained, what is the difference between open and closed models, what do training and answering cost, what is an API, how do MCP and agent skills extend an agent, and what is a System One model?
I · Agents	1–3	What is an agent, how does it take actions, and how do I keep that honest?
II · Machine learning	4–5	How do I turn a business question into a model I can trust?
III · Systems and decisions	6–7	How do I run this in production, and when should I not?
IV · Finance	8–9	What does all of this look like when the domain is markets?

Before session 1

Read the firm page, then the setup page. It takes ten minutes and gets you a working Colab with an API key.

Start here

• The firm: Champaign Capital Research
• Setup: Python, keys, Colab
• Schedule and rules

Part 0 · Foundations for AI

• 0. How Large Language Models Work
• 0A Transformers and Mixture of Experts
• 0B How a Model Is Trained
• 0C Pre-, Mid- and Post-Training
• 0D Open-Source and Closed-Source Models
• 0E Training and Inference Compute
• 0F What an API Is
• 0G Agent Skills and MCP
• 0H System One Models

Part I · Agents

• 1. Introduction to AI Agents
• 2. Prompt and Context Engineering
• 3. Memory Retrieval and RAG

Foundations for ML

• Foundations for ML

Part II · Machine learning

• 4. ML Foundations I
• 5. ML Foundations II

Part III · Systems and decisions

• 6. Agent Systems
• 7. When to Use ML, Agents, or Neither

Part IV · Finance

• 8. Financial Markets and AI as an Investment Theme
• 9. AI for Financial and Investment Research