Prompt Engineering Deep Dive
Power users and builders who want systematic prompting skills, beyond module 02
Prerequisites: Module 02 or equivalent. For the evaluation block: comfort with spreadsheets or notebooks.
Last verified 2026-07-28
Learning objectives
- Master the full prompting toolkit: structure, roles, separating data from instructions, output formatting, step-by-step thinking, examples
- Reduce hallucinations with explicit techniques (permission to say "I don't know", evidence-first answers)
- Build complex prompts for real industry cases (legal, financial, support)
- Write a basic prompt evaluation so quality is measured, not felt
- Know the advanced patterns by name: prompt chaining, tool use, retrieval
What's in this module
8 lessons · 7 exercises · ~8 min reading
Starts withStructure before cleverness~1 minExecutive summary
You will learn the complete prompting toolkit, from structure and roles to hallucination control and complex industry prompts, plus how to measure prompt quality with evaluations. End state is one production-grade prompt for your own domain, scored against a 5-case evaluation you wrote. After this module you can build prompts systematically instead of by feel, prove one version beats another with data, and recognize when the right tool is a skill or an agent rather than a prompt.
Sources
- Anthropic, "Prompt engineering overview", the model-independent techniques, and the index for the rest of this list: https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/overview
- Anthropic, "Prompting Claude Opus 5", over-verification, subagent delegation, response length, and what to strip from prompts written for older models: https://platform.claude.com/docs/en/build-with-claude/prompt-engineering/prompting-claude-opus-5
- Anthropic, "Effort", the primary control for token cost and latency: https://platform.claude.com/docs/en/build-with-claude/effort
- Anthropic's interactive prompt engineering tutorial: https://github.com/anthropics/prompt-eng-interactive-tutorial
- Anthropic courses repo (real-world prompting, prompt evaluations): https://github.com/anthropics/courses
Prefer the live room?
This module also runs inside our in-person bootcamps and workshops in Copenhagen.
More in Track F
Claude Code for Developers
How to run Claude Code as a professional engineering tool: operating manuals, subagents, hooks, skills, MCP, and git/CI workflows.
Agent Harness Engineering
How agents actually work under the hood: the model supplies intelligence, and everything else is a harness you can build.
Agentic Futures Briefing
Where agentic engineering is heading across three horizons: what's standard practice now, what arrives within a quarter, and what's forming six months out.