AI agent glossary
Context engineering
The practice of deciding exactly what information goes into a model's context window at each step so an agent has what it needs and nothing that distracts it.
Prompt engineering focuses on how you phrase a single request. Context engineering takes a wider view: everything the model sees on a given turn, including the system prompt, tool definitions, retrieved documents, memory, conversation history, and tool results.
For agents this matters enormously, because context changes every step. Good context engineering means loading information just in time, keeping tool outputs concise, summarizing or clearing old history, and moving durable knowledge into files or skills the agent can pull in when needed.
The guiding question is simple: given what the model sees right now, could a smart new colleague do this step well? If not, something is missing, or something irrelevant is crowding out what matters.
Learn more in the course
Related terms
Go from definitions to doing
Learn to use and build AI agents in 65 practical lessons.