Prompt Engineering – 100+ Lab Exercises (Basic, Intermediate, Advanced)

🔰 Basic Level (30+ Exercises)

 

Goal: Understand foundational concepts, experiment with basic prompt formats, and explore prompt behaviors.

 

Foundations of Prompt Engineering

  • Introduction to Large Language Models (LLMs).

  • What is Prompt Engineering? Key terminology.

  • Understand zero-shot, one-shot, and few-shot prompting.

  • Introduction to prompt templates and input-output structure.

  • Explore model temperature and top-p settings.

Basic Prompt Experiments

  • Write simple factual prompts (e.g., “What is a computer?”).

  • Generate lists (e.g., “List 5 countries in Asia”).

  • Reformat text (e.g., “Convert this to title case”).

  • Translate between languages using prompts.

  • Ask the model to summarize a paragraph.

Output Formatting

  • Ask LLM to return answers in Markdown format.

  • Prompt the model to output JSON data.

  • Request numbered lists or bullet points.

  • Extract structured data from unstructured input.

  • Wrap responses in HTML for integration with UI.

Basic Roles & Context Setting

  • Set persona: “You are a teacher explaining…”.

  • Impose tone: formal vs. casual vs. humorous.

  • Specify output audience (e.g., for 6th graders).

  • Ask for step-by-step reasoning.

  • Request analogy-based answers.

 

🚀 Intermediate Level (40+ Exercises)

 

Goal: Craft intelligent, reusable prompt systems and understand fine control over output using context, constraints, and chaining.

 

Prompt Patterns & Strategies

  • Chain-of-Thought prompting.

  • ReAct (Reasoning + Acting) prompting.

  • Tree-of-Thought prompting.

  • Role-play based prompt flows.

  • Reflection-based self-improvement prompts.

Prompt Engineering for Tools & APIs

  • Design prompts for code generation (e.g., Python, JS).

  • Create structured SQL queries from natural language.

  • Generate REST API documentation from prompt.

  • Debug code using error-prompt explanation.

  • Prompt for test case generation.

Contextual Prompting

  • Supply background context in prompts.

  • Use delimiter markers (""", <<< >>>) for large inputs.

  • Reference prior responses in chained prompts.

  • Control memory with system-level instructions.

  • Inject temporal or location-based constraints.

Industry-Specific Prompting

  • Finance: Extract financial data from news articles.

  • Healthcare: Summarize patient notes.

  • Law: Rephrase legal documents.

  • Education: Generate custom quizzes.

  • HR: Draft emails and performance reviews.

 

🧠 Advanced Level (40+ Exercises)

 

Goal: Create production-grade prompt systems, integrate with APIs and agents, and apply prompt engineering for autonomous systems.

 

Advanced Prompt Design & Optimization

  • Token management and compression strategies.

  • Use of embeddings in retrieval-augmented generation (RAG).

  • Convert long documents into vectorized chunks for input.

  • Dynamic prompt injection with APIs (LangChain, LlamaIndex).

  • Prompt optimization using eval()-based testing.

Agents & Autonomous Prompt Systems

  • Prompt chaining using LangChain agents.

  • Build multi-step tool-using agents (e.g., web + code).

  • Construct tools with function calling.

  • Use prompt routers for context switching.

  • Multi-agent communication (e.g., teacher ↔ student).

Prompt Evaluation & Debugging

  • Evaluate prompt accuracy, safety, and hallucination.

  • Identify prompt-induced biases and reduce them.

  • Use OpenAI’s evals framework or custom test sets.

  • Implement A/B testing for prompts.

  • Log and version-control prompts with metadata.

Ethical & Secure Prompting

  • Prevent jailbreaks in prompts.

  • Sanitize user input in public prompt forms.

  • Control offensive output with constraints.

  • Add moderation layers using prompt validation.

  • Use guardrails to block harmful queries.

 

Capstone Projects

 

  • Design a chatbot that provides legal advice using prompt chaining.

  • Build a content generation pipeline (SEO blogs, emails).

  • Create a self-improving agent with feedback loops.

  • Implement an AI tutor with memory-based prompts.

  • Develop a code review assistant with prompt-based correction.

 

✅ Tools & Libraries

 

  • LangChain, LlamaIndex, PromptLayer

  • OpenAI Playground, Claude Console, Gemini Studio

  • Gradio, Streamlit, Chainlit

  • Replit, Jupyter Notebooks

  • LangSmith (for logging, tracing)

 

📚 Resources

 

  • OpenAI Cookbook, Anthropic Prompt Guides

  • Papers like “Prompt Programming for Large Language Models”

  • OpenPrompt, DeepEval, PromptSource

  • Blogs: Latent Space, Weights & Biases, OpenAI Engineering

Download Syllabus 👇
Scroll to Top