Course

Prompt Engineering

Zero-shot, few-shot, chain-of-thought, role framing, retrieval grounding, self-consistency, context limits, and prompt chaining, taught with real model transcripts and tied directly back to the attention and sampling mechanics from Zero to GPT.

Lesson 1

How Prompts Actually Work

Every technique in this course is really just a decision about what tokens to put in the context window.

Start

Lesson 2

Zero-Shot vs. Few-Shot Prompting

Sometimes examples fix correctness. More reliably, they fix format.

Start

Lesson 3

Chain-of-Thought Prompting

Asking a model to "think step by step" gives it more of its own tokens to attend to, and that's not a metaphor.

Start

Lesson 4

System Prompts & Role Framing

The same mechanism the AI Security course used to explain prompt injection, now used on purpose.

Start

Lesson 5

Grounding with Retrieved Context

Every other lesson assumed the model already knew what it needed, this one is about the tokens you have to hand it when it doesn't.

Start

Lesson 6

Self-Consistency & Structured Output

Sampling more than once, and asking for exactly the shape you need, both callbacks to Zero to GPT's sampling lesson.

Start

Lesson 7

Context Limits in Practice

Every technique in this course shares one hard constraint that hasn't come up directly until now, the window has a fixed size.

Start

Lesson 8

Prompt Chaining

The capstone: breaking a complex task into a sequence of focused calls, each with its own clean context window.

Start