Zenaique
Glossary · Tooling

Context Engineering

Also known as: Context management

Deciding what makes it into the model's context window (prompt + RAG + history + memory), and why.

The discipline of optimizing what information ends up in an LLM's context window (including prompt structure, retrieval, summarization, history management, and memory) to maximize task quality within token and latency budgets.

In practice

The umbrella discipline above prompt engineering and RAG. Senior LLM app design interviews increasingly frame the problem this way.

How it compares

Prompt engineering tunes the static prompt; context engineering manages everything that flows in dynamically.

Related topics

Related terms