Pathmaker
Prompt Engineering Mastery
Master the art and science of prompt engineering to get exceptional results from AI models. Learn techniques used by top practitioners to build reliable AI workflows.
About This Course
Prompt engineering is the critical skill for working effectively with AI in 2026. This hands-on course teaches you systematic approaches to crafting prompts that consistently produce high-quality results.
You'll learn from real-world examples, build a personal prompt library, and understand how to optimize prompts for different models and use cases.
What You'll Learn
- 1Write effective prompts that consistently achieve desired outcomes
- 2Understand prompt patterns and when to apply each technique
- 3Debug and iterate on prompts systematically
- 4Build reusable prompt templates for common tasks
- 5Optimize prompts for cost, speed, and quality
- 6Chain prompts together for complex workflows
Week by Week
Foundations of Effective Prompting
Start with the fundamentals. Learn how language models actually process prompts, what makes them succeed or fail, and the core patterns that professionals use daily.
Explore this week's curriculum
Key concepts
Zero-shot vs Few-shot: Zero-shot provides no examples, few-shot includes 2-5 examples of desired input/output pairs to guide the model
Context Window: The amount of text (prompt + response) a model can process at once, measured in tokens
Temperature: Parameter controlling randomness vs consistency in responses (0 = deterministic, 1 = creative)
System Prompt: Instructions that set the AI's role, behavior, and constraints for the entire conversation
Studio activities
Build Your First Prompt Library
Create 10 prompts for tasks you do regularly (summarizing, analyzing, writing, etc.) and test variations to find what works best.
- List 10 tasks you want AI help with
- Write a basic prompt for each task
- Test each prompt 3 times with different inputs
- Document what works and what fails
- Create one improved version of each
- Save your best prompts in a personal library
Success criteria
A collection of 10 tested, working prompts with notes on when to use each
Prompt Pattern Recognition
Analyze successful prompts from the community to identify patterns and techniques.
- Find 20 high-quality prompts from sources like PromptBase, Reddit, or Twitter
- Categorize them by use case
- Identify common structural patterns
- Note specific techniques (examples, constraints, formatting)
- Extract reusable patterns for your own prompts
Success criteria
Document 5-7 prompt patterns with examples of when to use each
Reflection
- What surprised you most about how AI responds to different prompts?
- Which prompt pattern felt most natural to you?
- What tasks are you most excited to automate with better prompts?
Study guidance
- Test every prompt at least 3 times to check consistency
- Keep a prompt journal to track what works
- Start simple, then add complexity only when needed
Discuss with Professor Reed
Choose a question to discuss with Professor Reed. Copy it and paste it into your conversation with him. The copied text includes this week's lesson and assignment; add your own questions or explain where you are stuck.
Why do some prompts work better than others?
How do I make my prompts more specific without being too rigid?
What's the difference between prompting ChatGPT vs Claude vs other models?
What surprised you most about how AI responds to different prompts?
Which prompt pattern felt most natural to you?
What tasks are you most excited to automate with better prompts?
Ask me one question at a time to check my understanding, then explain any gaps.
Help me plan this assignment without doing the work for me.
Advanced Techniques & Prompt Patterns
Level up with advanced techniques: chain-of-thought, role prompting, constraint setting, and more. Learn when and how to apply each method for maximum effectiveness.
Explore this week's curriculum
Key concepts
Chain-of-Thought (CoT): Asking the model to show its reasoning step-by-step before giving the final answer
Role Prompting: Assigning the AI a specific role or persona to shape its responses (e.g., 'You are an expert financial analyst')
Constraints & Guardrails: Explicit rules and boundaries to keep AI responses on track and safe
Output Formatting: Specifying exact structure for responses (JSON, markdown, tables, etc.)
Studio activities
Chain-of-Thought Mastery
Take complex reasoning tasks and compare direct prompting vs chain-of-thought approaches.
- Choose 5 complex problems (analysis, decision-making, etc.)
- Write direct prompts for each
- Rewrite with chain-of-thought instructions
- Test both approaches 3 times each
- Compare accuracy and quality
- Document when CoT adds value vs when it's unnecessary
Success criteria
Clear understanding of when chain-of-thought improves results
Build a Structured Output System
Create prompts that reliably produce structured data (JSON, CSV, etc.) for a real use case.
- Define a task needing structured output (e.g., extracting data from text)
- Design your desired output schema
- Write a prompt with clear formatting instructions
- Test with varied inputs
- Add error handling and validation
- Document the final prompt with examples
Success criteria
Prompt that produces valid structured output 95%+ of the time
Reflection
- Which advanced technique had the biggest impact on your results?
- What challenges did you face getting consistent outputs?
- How do you balance prompt complexity vs simplicity?
Study guidance
- Chain-of-thought works best for reasoning and math problems
- Be specific about constraints - vague rules get ignored
- Test output formats with edge cases
Discuss with Professor Reed
Choose a question to discuss with Professor Reed. Copy it and paste it into your conversation with him. The copied text includes this week's lesson and assignment; add your own questions or explain where you are stuck.
When should I use chain-of-thought vs direct prompting?
How do I get consistent structured outputs?
What makes a good role prompt vs a bad one?
Which advanced technique had the biggest impact on your results?
What challenges did you face getting consistent outputs?
How do you balance prompt complexity vs simplicity?
Ask me one question at a time to check my understanding, then explain any gaps.
Help me plan this assignment without doing the work for me.
Prompt Optimization & Debugging
Learn systematic approaches to debugging bad prompts, optimizing for cost and speed, and A/B testing different approaches. Make your prompts production-ready.
Explore this week's curriculum
Key concepts
Prompt Debugging: Systematic process for identifying why prompts fail and fixing them
Token Optimization: Reducing prompt length to lower costs while maintaining quality
Prompt Versioning: Tracking changes to prompts over time to improve iteratively
A/B Testing: Comparing multiple prompt variations to find the best performer
Studio activities
Token Diet Challenge
Take your longest prompts and reduce them by 50% while maintaining output quality.
- Select 3 of your longest prompts
- Establish baseline quality with 10 test runs each
- Remove unnecessary words and repetition
- Test the shortened version
- If quality drops, add back only essential parts
- Measure final token savings and quality impact
Success criteria
50% token reduction with <10% quality loss
Prompt A/B Testing Framework
Create a systematic process for testing and improving prompts.
- Choose a prompt to optimize
- Create 3-5 variations with different approaches
- Define success metrics (accuracy, tone, format, etc.)
- Test each variation with 20+ diverse inputs
- Score results objectively
- Document winning approach and why it worked
Success criteria
Measurable improvement with documented evidence
Reflection
- What was your biggest 'aha moment' during optimization?
- How much can you reduce prompt length before quality suffers?
- What metrics matter most for your use cases?
Study guidance
- Always test prompt changes with the same inputs for fair comparison
- Shorter prompts usually work just as well as verbose ones
- Version control your prompts like code
Discuss with Professor Reed
Choose a question to discuss with Professor Reed. Copy it and paste it into your conversation with him. The copied text includes this week's lesson and assignment; add your own questions or explain where you are stuck.
How do I debug a prompt that sometimes works and sometimes fails?
What's the best way to reduce token usage without losing quality?
How do I know when a prompt is 'good enough'?
What was your biggest 'aha moment' during optimization?
How much can you reduce prompt length before quality suffers?
What metrics matter most for your use cases?
Ask me one question at a time to check my understanding, then explain any gaps.
Help me plan this assignment without doing the work for me.
Advanced Workflows & Real-World Applications
Put it all together. Build complex multi-step workflows, integrate prompts into applications, and create your own prompt-based tools that solve real problems.
Explore this week's curriculum
Key concepts
Prompt Chaining: Using the output of one prompt as input to the next to accomplish complex tasks
Dynamic Prompts: Prompts that adapt based on variables, user input, or previous outputs
Prompt Templates: Reusable prompt structures with placeholders for different inputs
Error Recovery: Strategies for handling when prompts produce unexpected or invalid outputs
Studio activities
Build a Prompt-Powered Tool
Create a complete workflow or tool using everything you've learned.
- Identify a real problem you want to solve
- Design a multi-step workflow to solve it
- Create prompts for each step
- Connect the prompts into a chain
- Add error handling and validation
- Test with real-world inputs
- Document your tool and share it
Success criteria
Working tool that solves a real problem end-to-end
Publish Your Prompt Library
Package your best prompts and techniques into a shareable resource.
- Review all prompts you've created
- Select your top 20 most useful ones
- Write clear documentation for each
- Organize by category and use case
- Add examples and tips
- Share on GitHub, Notion, or your blog
Success criteria
Published collection of documented, tested prompts
Reflection
- What's the most valuable prompt or workflow you created?
- How has prompt engineering changed how you work with AI?
- What will you build next with these skills?
Study guidance
- Start with the end output and work backwards to design chains
- Always validate intermediate outputs before passing to next step
- Build error handling from the start
Discuss with Professor Reed
Choose a question to discuss with Professor Reed. Copy it and paste it into your conversation with him. The copied text includes this week's lesson and assignment; add your own questions or explain where you are stuck.
How do I design effective prompt chains?
What's the best way to handle errors in multi-step workflows?
How do I make prompts that work for different users/inputs?
What's the most valuable prompt or workflow you created?
How has prompt engineering changed how you work with AI?
What will you build next with these skills?
Ask me one question at a time to check my understanding, then explain any gaps.
Help me plan this assignment without doing the work for me.
Topics Covered
More in This Track
Designing for Delegated Experiences
Design how people entrust decisions and actions to people, services, and agents.
Algorithms and the Online Experience
Investigate how ranking, recommendation, advertising, and moderation mediate everyday online life.
Design, Technology, and AI Systems
How designers can understand AI products, software ecosystems, automation, interfaces, and workflow tools — without becoming purely technical.
Ready to begin?
Start your personalized curriculum today — free to try, no credit card needed.