Prompt Engineering

Full definition

Effective prompts typically combine: a clear role and task, relevant context, illustrative examples (few-shot), explicit output format, and edge-case rules. Advanced techniques include chain-of-thought reasoning, ReAct (reason + act for tool use), and self-consistency (sampling multiple answers and majority-voting). Prompt engineering is the first thing every team should master before considering fine-tuning.

Why it matters

Prompt engineering converts a generic foundation model into a specific, dependable production component. Rigorous prompt engineering, plus eval-driven iteration, can match fine-tuned model quality at zero training cost — and is reversible the next time a better foundation model ships.

Example

A support team rewrites their classification prompt to include 5 examples, an explicit JSON output schema, and a "if uncertain, return UNKNOWN" rule. Accuracy rises from 78% to 94%; downstream parsers stop crashing.

Related terms

• Large Language Model (LLM)
  A Large Language Model (LLM) is a deep neural network — almost always a transformer — trained on hundreds of billions to trillions of words to predict the next token, and to generate, summarize, translate, or reason over text.
• Few-shot / Zero-shot Learning
  Few-shot learning is the ability of a model to perform a new task given only a handful of examples at inference time; zero-shot learning is the same idea with no examples — the model relies entirely on its pretraining and the natural-language instruction.
• Chain-of-Thought (CoT)
  Chain-of-Thought (CoT) is a prompting and training technique in which a language model is encouraged to reason step-by-step before producing its final answer — improving accuracy on math, logic, and multi-hop reasoning by 10-40 percentage points on standard benchmarks.
• Context Window
  A context window is the maximum amount of text — measured in tokens — a language model can read and reason over in one inference call, equivalent to the model's working memory for that turn.
• Retrieval-Augmented Generation (RAG)
  Retrieval-Augmented Generation (RAG) is a technique where a language model retrieves relevant documents from an external knowledge base before generating a response, reducing hallucination by 30-60% and enabling citation of sources.