Learn to work effectively with AI coding assistants. This course teaches the practical skills of prompting, context management, and AI-assisted workflows.
Option A: Download ZIP (easiest)
- Click the green "Code" button above
- Select "Download ZIP"
- Unzip to a folder on your computer
Option B: Clone with Git
git clone https://github.com/Aurite-ai/ai-collaboration-course.gitOnce you have the files locally, you can use them with Kahuna:
- Open the course folder in VS Code
- Share assignment files with Kahuna as context
- Follow the assignment instructions
| Session | Topic | Status |
|---|---|---|
| Session 1 | Effective Prompting | ✅ Available |
| Session 2 | The Two Modes of Working | ✅ Available |
| Session 3 | The Generate-Critique Cycle | 🔜 Coming Soon |
| Session 4 | TBD | 📅 Planned |
| Session 5 | TBD | 📅 Planned |
What you'll learn:
- What an LLM actually is (text in, text out)
- Why "everything is text" matters for prompting
- Two types of prompts: literal instructions vs. decision-making
- The G/R/S/O framework for complex tasks
Materials:
- 📊 Slides — Marp format (readable as markdown)
Assignments: None (conceptual foundation session)
What you'll learn:
- The fundamental difference between exploration and execution
- How G/R/S/O prompts produce design documents and plans
- Why documents bridge the two modes
- Context accumulation and how to manage it
Materials:
- 📊 Slides — Marp format
- 📝 Assignments — Practice the two-mode workflow
Recommended: Complete Assignment 1 before Session 3.
The slides are in Marp format — a markdown-based presentation tool.
To read them:
- Open the
.mdfile directly — the content is readable as markdown - Speaker notes appear in
<!-- comment -->blocks
To view as slides (optional):
- Install the Marp for VS Code extension
- Open the file and click the preview icon
You don't need the slide view — the markdown is the content.
Additional materials for curious students:
- Assignment Design Framework — How the course assignments are designed (three difficulty axes)
- Basic comfort with coding (any language)
- Python and pandas for assignments (data analytics focus)
- Access to an AI coding tool (Kahuna recommended, or Claude Code, Cursor, etc.)
This course was developed for USC data analytics graduate students learning to work with AI coding assistants. The materials are designed for:
- Students new to AI coding tools
- People who've tried AI tools but found them inconsistent
- Anyone who wants to level up their AI collaboration skills
The assignments use data analytics examples (CSV processing, data cleaning) but the skills transfer to any domain.
MIT License — feel free to use, share, and adapt these materials.
Learn to work WITH the AI, not fight against it.