Context Engineering

For people who have already written the clever prompt and hit its ceiling. This course treats the context window as an engineering surface: what you put in it, in what order, from where, at what cost, and how you prove a change was an improvement. Nine lessons covering token budgets, the system prompt boundary, forced schemas, retrieval against long context, chunking, prefix caching, positional attention, untrusted input, and shipping prompts you can version and roll back. Vendor-neutral, with real numbers.

1. 1The window is a budget, not a container8 minEvery token costs money, latency and attention, so treat the window as spend, not storage.
2. 2What belongs in the system prompt8 minStable and yours goes on top; variable or untrusted goes below, always.
3. 3Structured output, and what a schema does not buy you9 minA schema constrains the shape of an answer, never its truth, and its field order is a prompt.
4. 4Retrieval versus long context: the real trade9 minCorpus size, permissions and freshness pick the architecture; window size only changes the budget.
5. 5Chunking that does not destroy meaning9 minRetrieve the unit a human would quote, and carry down the context that made it mean what it means.
6. 6Caching a stable prefix8 minCaching matches from the first token forward, so anything variable near the top throws the whole discount away.
7. 7Ordering, and the fact that models skim8 minPosition changes whether a fact gets used, so measure the position curve instead of trusting rank order.
8. 8When untrusted text enters the context9 minYou cannot stop a model from reading an instruction, so limit what it is able to do about one.
9. 9A prompt you can version, test and roll back10 minVersion the whole call and log its hash, or you will not be able to explain your own regressions.