Context Engineering for AI-Assisted Coding
Published by O'Reilly Media, Inc.
Strategies for working with code using Claude, Cursor, or any other AI coding assistant
What you’ll learn and how you can apply it
- Learn context engineering strategies for using Claude Code to build, analyze, and maintain code
Course description
This comprehensive, hands-on course with Chelsea Troy introduces software developers to the capabilities of AI coding assistants, specifically Cursor and Claude, in the context of coding workflows. LLMs produce more accurate and useful answers when engineers practice judicious context management.
You’ll start by establishing your intuition for how LLMs use context and learn about what both model designers and model deployers can do to change that. Then you’ll explore how that specifically affects engineers in the process of navigating code bases with LLM assistance. You’ll also pick up both code characteristics and LLM usage practices to improve your results. Finally, you’ll scope the sorts of situations in which LLM tools are most useful.
Whether you’re working within a proprietary code base or experimenting with open source repositories, this course equips you with the tools and understanding to anchor Claude, Cursor, or any other AI coding assistant as a collaborative partner in your software development process.
This live event is for you because...
- You’re a software engineer.
- You work with tools like Claude or Cursor.
- You want to become more effective at using these tools in coding work.
Prerequisites
- Have your AI coding tool of choice installed
Recommended preparation:
- Claude Code ready to go, either in Cursor or in the VS Code extension
Recommended follow-up:
- Take Agentic Coding with Claude Code (live course with Ken Kousen)
- Take Comparing Cursor, Copilot, and Windsurf (live course with Sergio Pereira)
Schedule
The time frames are only estimates and may vary according to how the class is progressing.
Introduction (10 minutes)
- Presentation: Who I am, who you are, and what we’re doing here
Context management (45 minutes)
- Presentation: How context works in LLM enabled tools; the /clear, /compact, /exit, and /resume commands and when to use them; @-files for narrowing down context; putting a cursor in a file
- Hands-on exercise: /clear your Claude; hit /resume and see what it thinks your prior conversation was
- Q&A
Commands and their role in context management (20 minutes)
- Presentation: Options for saving to memory and how to use them; terminal setup, slash commands, ! for bash commands, and the # and /memory commands
- Hands-on exercise: Save something into the Claude file using the hash command
- Break
Characteristics of an AI-compatible code base (15 minutes)
- Presentation: What makes LLM products effective on code generally, and what aspects of specific code bases improve your likelihood of finding utility with the tool?
- Q&A
Summarizing the code base with context documents (15 minutes)
- Presentation: Why these pieces matter to you, and why they matter to the tool; editing the Claude file manually; how the file is used by Claude
- Hands-on exercise: Initialize Claude or Cursor on a code base
Plan mode (60 minutes)
- Presentation: What’s actually happening when you do plan mode; what it means for your context profile
- Demonstration: Running a plan
- Hands-on exercise: Construct and edit a plan to make a change in your code base
- Group discussion: Does the Claude plan mention any steps that you would have skipped?; What edits would you make to the plan?
- Break
Think mode (20 minutes)
- Presentation: How to activate think mode and what it’s doing under the hood; advice for using and not using think mode
- Q&A
Wrap-up (5 minutes)
Your Instructor
Chelsea Troy
Chelsea Troy leads the machine learning operations team at Mozilla. She also teaches in the Master’s Program in Computer Science at the University of Chicago. Her online workshop, Fundamentals of Technical Debt, is available On Demand through the O’Reilly platform, and she also gives live courses about machine learning, large language models, and product thinking.