Glossary · 3 · Working with models: prompts and context
Context engineering
Also known as: Context design
Context engineering is the discipline of deciding what information enters a language model’s context window for each step of a task — instructions, retrieved documents, examples, conversation history, tool results and memory — and in what form, so the model has exactly what it needs and nothing that misleads it.
- Expert
- Technical writers
- Technical project managers
- Developers
In one sentence
Context engineering explained: curating everything an AI model sees — documents, tools, memory — the expert skill behind reliable AI systems.
Example
Instead of pasting a whole manual, a documentation agent retrieves the two relevant topics, adds the product variant from metadata and a short summary of the conversation, then asks the model to answer.
Why it matters on your learning path
- Technical writers: The quality of context depends on the quality of content. Structured, metadata-rich content is the raw material of context engineering.
- Technical project managers: Most quality problems in AI features are context problems, not model problems. Budget for content preparation.
- Developers: Design retrieval, compaction and memory as first-class parts of the system; measure the effect of each context change.