Prompt engineering: get reliable answers from chat models
11 steps · about 7 h 30 min · Beginner
- Writing Effective AI Prompts for Business (guide, Google Workspace, 15 min)
A short, jargon-free start that introduces the four parts of a prompt: persona, task, context and format.
- Prompt engineering best practices for ChatGPT (article, OpenAI Help Center, 10 min)
Adds the two habits everything else builds on: be specific, then iterate on what the model gives back.
- Prompt-Engineering-Guide/guides/prompts-intro.md (guide, DAIR.AI, 20 min)
Shows how a prompt is built from elements such as instruction, context, input and output format, with before-and-after examples.
- Anthropic Prompt Engineering Interactive Tutorial (interactive, Anthropic (GitHub), 2 h)
Hands-on exercises let you practise clarity, roles, examples and step-by-step reasoning, and see the effect of each.
- ChatGPT Prompt Engineering for Developers (video, DeepLearning.AI (Andrew Ng, Isa Fulford), 1 h 40 min)
Teaches two core principles and an iterate-and-test loop through short lessons on summarizing, inferring, transforming and expanding text.
- Prompt engineering overview (doc, Claude Docs (Anthropic), 15 min)
Explains when prompting is the right fix and lists techniques ordered from most to least broadly effective.
- Prompt design strategies (doc, Google AI for Developers, 25 min)
A second vendor's view on structure, context, and splitting or chaining prompts, so you see which habits work across models.
- Prompt engineering (doc, OpenAI, 25 min)
Covers separating instructions from data with Markdown and XML tags, and versioning prompts like code.
- Prompt Engineering Guide (reference, DAIR.AI, 1 h)
Gives the named techniques in one place, such as few-shot, chain-of-thought and prompt chaining, plus guidance on reliability and bias.
- Prompting best practices (doc, Claude Platform Docs (Anthropic), 40 min)
A deeper reference on clarity, examples, structure and thinking, for tightening prompts you already use.
- Effective context engineering for AI agents (article, Anthropic, 20 min)
Closes by widening the view from the wording of a prompt to curating everything the model sees.