Prompt Engineering Lab

A course-to-code learning lab for studying prompt engineering through notes, experiments, and publishable artifacts.

Overview

This project follows a simple workflow: read, explain ideas in plain language, build small experiments, and turn the results into reusable learning materials.

Learning Roadmap

Chapter	Status	Notes	Notebook	One-line Takeaway
Introduction	✅ Done	notes/introduction.md	notebooks/01_introduction_practice.ipynb	Prompt engineering is the systematic design of inputs to reliably bring out LLM capability.
LLM Settings	✅ Done	notes/llm-settings.md	notebooks/02_llm_settings_practice.ipynb	Model behavior depends not only on prompts but also on decoding settings and model configuration.
Basics of Prompting	✅ Done	notes/basics-of-prompting.md	-	Clear instructions, context, and examples make prompts more reliable than raw text alone.
Prompt Elements	✅ Done	notes/prompt-elements.md	notebooks/03_prompt_elements_practice.ipynb	A strong prompt is built from reusable parts such as instruction, context, input data, and output indicator.
General Tips	✅ Done	notes/general-tips.md	notebooks/04_general_tips_practice.ipynb	Start simple, be specific, separate instructions clearly, and say what to do instead of only what not to do.
Zero-Shot	✅ Done	notes/zero-shot.md	notebooks/05_zero_shot_practice.ipynb	Zero-shot prompting asks the model to do a task directly without demonstrations, relying on its instruction-following ability.
Few-Shot	✅ Done	notes/few-shot.md	-	Few-shot provides labeled demonstrations so the model learns the task format and label space through in-context learning.
Chain-of-Thought	✅ Done	notes/chain-of-thought.md	-	CoT adds intermediate reasoning steps to demonstrations so the model learns how to solve problems, not just what format to answer in.
Meta Prompting	✅ Done	notes/meta-prompting.md	-	Meta Prompting uses abstract structural templates instead of content examples, shifting the work of pattern extraction from the model to the prompt writer.
Self-Consistency	✅ Done	notes/self-consistency.md	-	Self-Consistency samples multiple CoT reasoning paths and aggregates final answers by majority vote to improve reliability on deterministic reasoning tasks.
Tree of Thoughts	✅ Done	notes/tree-of-thoughts.md	-	Tree of Thoughts treats reasoning as a search tree over intermediate thoughts, enabling branching, evaluation, and backtracking.
RAG	⬜	-	-	-
ReAct	⬜	-	-	-
Adversarial Prompting	⬜	-	-	-
AI Agents	⬜	-	-	-

Project Structure

prompt-engineering-lab/
├── notes/
├── notebooks/
├── prompts/
├── projects/
├── .env.example
├── .gitignore
├── LEARNING_LOG.md
├── README.md
└── requirements.txt

Quick Start

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
cp .env.example .env
python -m ipykernel install --user --name prompt-engineering-lab
jupyter notebook

Learning Log

Session-based notes live in LEARNING_LOG.md. Chapter-level completion status stays in the roadmap above.

• Learning Log

References

• Prompt Engineering Guide
• Feynman Technique