Using models well · Part 3 — The AI Dictionary
Context engineering
Deliberately choosing what goes into the model's context — instructions, files, examples, tools — to get reliable results.
Prompt engineering is wording one request well. Context engineering is designing everything the model sees: which documents, which past messages, which tools, which project instructions (like a CLAUDE.md or AGENTS.md file), and what to leave out.
📝 Read the full lesson notes
Context engineering
Deliberately choosing what goes into the model's context — instructions, files, examples, tools — to get reliable results.
Prompt engineering is wording one request well. Context engineering is designing everything the model sees: which documents, which past messages, which tools, which project instructions (like a CLAUDE.md or AGENTS.md file), and what to leave out.
API
The programmatic way to use a model from your own code or app, usually paid per token.
Chat apps are for people; APIs are for software. Through an API, developers send prompts and receive responses in code, choose the exact model, set temperature, enforce structured output and pay per million tokens used.
Latency
How long you wait for a response — time to first token plus generation speed.
Latency is the delay between hitting Enter and seeing a useful answer. Bigger models and reasoning modes are slower; smaller or distilled models are faster. For voice assistants and autocomplete, latency matters more than raw smarts.
Part of 📖 The AI Dictionary in the free AI Bootcamp.