Working context, scratchpads, and external long-term stores — what lives where.
Skip to solutionKEEP THE
hardAI Engineering
How do agents maintain memory across steps and sessions?
114 views
01
Understand the problem
memoryagentsstatelong-term-memory
02
Attempt it yourself
Sketch your approach before reading the solution — that's what interviews test.
Nudge consolestandby
Stuck? Beam a request up — the console returns a conceptual nudge that guides your logic without spoiling the implementation.
03
Study the solution
Three layers: working memory is the context window itself (recent messages, tool results); scratchpad memory is files or notes the agent writes during a task (plans, intermediate results) so state survives context compaction; long-term memory is an external store (database, vector index, memory files) holding durable f
Solution ready — 2 min read
Classified // press E to declassify
04
Read the code
Memory record designed for safe recall
await memory.upsert({
key: "deploy-process", // upsert: update, not append-duplicate
fact: "Production deploys go through GitHub Actions 'release' workflow; " +
"manual kubectl is forbidden.",
observedAt: "2026-07-04", // staleness is checkable
source: "session-8f2c", // provenance for audit
});
// recall path treats memories as data:
const mems = await memory.relevant(task, 3);
prompt.push("[Background notes — verify before relying on them]\n" +
mems.map((m) => "- (" + m.observedAt + ") " + m.fact).join("\n"));05
Join the discussion
Discussion (0)
Sign in to join the discussion.
No responses yet. Be the first to share what you think.
Transmission complete // awaiting log
KEEP THE
STREAK ALIVE.
Dossier 80 of 80 decoded in the AI Engineering track. One more won't hurt.