What is Context Engineering

Definition

Context engineering is the practice of deliberately assembling everything a language model sees at inference time, instructions, retrieved data, examples, history, and tools, so the model has exactly the right information to perform a task well, extending prompt engineering to the full context window.

« Back to Glossary Index

Optimises the whole context, not just the prompt wording
Supplies relevant retrieved data so answers are grounded and accurate
Manages limited context space by prioritising what matters most
Improves reliability of complex, multi-source AI workflows

Real World Example

An engineer building an AI assistant carefully curates the context for each query, including the most relevant retrieved documents, prior conversation, and tool definitions, so the model consistently produces grounded, on-task answers.

FAQs

How does context engineering differ from prompt engineering?

Prompt engineering focuses on instruction wording, while context engineering manages the entire set of information, retrieved data, history, and tools, given to the model.

Why is context engineering important?

Because models reason only over what is in their context, assembling the right information is often decisive for output quality.

What is the challenge in context engineering?

Fitting the most relevant information into a limited context window while avoiding noise that degrades the model's focus.