Course
Agents That Remember: Context Engineering & Memory
7 lessons
An interactive, 15-minute distillation of Anthropic engineer Lamis Mukta's talk on context engineering and agent memory — how agents autonomously manage their own memory so their capability compounds ("a system that prompts itself"), and how to make that work in production. Covers the CLAUDE.md → memory-tools → skills → file-system evolution, progressive disclosure, versioning + hashing for concurrency, and read-only vs read-write permissioning. Predict, drag, and decide your way through.
What you’ll learn
1.Why Intelligence Doesn't Compound
A smarter model doesn't automatically make a better agent. Context — memory of your work — is the multiplier, and it has evolved fast: from one markdown file to autonomous file-system memory.
3 lessons
2.The Memory Playbook
The state of the art in two moves: progressive disclosure (deep detail without context bloat) and modeling memory as a plain file system the agent searches itself — Format, Reading, Writing.
2 lessons
3.Productionizing Memory
The neat idea meets reality: many agents, long runs, shared files. Solve concurrency with versioning + hashing, set read-only vs read-write permissions, keep memory portable — and watch capability compound.
2 lessons
How this course works
This is an interactive course, not a video playlist. Lessons mix short explanations with hands-on activities — quizzes, drag-and-drop exercises, branching scenarios and step-ordering challenges — and you have to complete each activity before moving on, so you practise rather than skim. Your progress is saved as you go, so you can stop after any lesson and pick up where you left off.
Topics covered
- ai
- agents
- context-engineering
- memory
- anthropic
- interactive