AI Agents · TRACK 04/05
Planning & Memory
How agents break down tasks, manage context, and remember across sessions.
// THE TRACK
01 · START HEREAgent PlanningSee how agents turn a vague goal into an ordered list of executable steps — and what they do when the plan stops matching reality.BEGINNERTypes of MemoryUnderstand the four memory types that let agents remember facts, past conversations, and learned skills — and how each one is stored and retrieved.INTERMEDIATEShort-Term vs Long-TermUnderstand the difference between what an agent holds in its context window and what it can recall across sessions.BEGINNERLong-Term MemoryCompare the three dominant approaches to persistent agent memory and learn which storage shape fits which kind of recall.INTERMEDIATEContext CompactionLearn how agents summarize their own history to stay under the context limit without forgetting what matters mid-task.INTERMEDIATEReflection & Self-CorrectionLearn how an agent can grade and revise its own work, the patterns that make self-correction reliable, and when reflection just wastes tokens.INTERMEDIATEAgent Stop ConditionsUnderstand the conditions that end an agent's loop — completion signals, step and budget limits, and stuck-detection — and why every agent needs them.BEGINNERScratchpad & Working MemoryLearn what an agent's scratchpad is, how it serves as working memory within a single task, and why it differs from both context and long-term memory.BEGINNERMem0You will understand what Mem0 does as a dedicated memory layer, how it persists and retrieves personalized facts across sessions, and where it fits in an agent's memory stack.INTERMEDIATE