Series · Builders · Step 2
Prompting & Context
A good prompt is closer to a formal letter than to a quick text message — and the real unit of control is not the prompt alone but the whole context the model sees. This series works up from the basics of prompting to context engineering: the standard prompting patterns, how to work with an LLM through its API, and how an agent carries memory across a conversation. The focus is on the underlying principles that apply to any model.
5 pieces, in order
-
Part 1: Prompt engineering from zero: write a formal letter, not a text message
Step 2 · 6 min -
Part 2: Zero-shot, few-shot, or chain-of-thought: picking the right technique
Step 2 · 5 min -
Part 3: From prompt engineering to context engineering
Step 2 · 4 min -
Part 4: Working with LLM APIs: first calls, tokens, and structured output
Step 2 · 6 min -
Part 5: How agents remember: memory and knowledge representation
Step 2 · 6 min