## TL;DR
Prompt engineering is the practice of designing text inputs to elicit desired outputs from language models. It has evolved from simple instruction writing to systematic techniques for reasoning, constraint specification, and multi-step task decomposition.

## Core Explanation
Key techniques: zero-shot (instruction only), few-shot (instruction + examples), chain-of-thought (step-by-step reasoning), tree-of-thoughts (exploring multiple reasoning paths), ReAct (reasoning + action interleaving), and structured output formatting (JSON mode, function calling).

## Detailed Analysis
Prompt templates with variables enable dynamic prompt construction for production applications. Retrieval-augmented generation (RAG) adds context from external knowledge bases. Constitutional AI uses principle-based prompts to steer model behavior toward safety. DSPy (Stanford, 2024) treats prompts as optimizable parameters.

## Further Reading
- OpenAI: Prompt Engineering Guide
- Anthropic: Prompt Library
- Prompting Guide (DAIR.AI)