Learning path
Agent Memory and State
The context window, long-term memory and summaries, and vector memory with embeddings, with animated films and runnable Python. Part of the Agentic AI Engineer premium track.
-
1
Short-Term Memory: the Context Window
An agent's working memory is its context window. Measure what fills it, keep it lean with sliding windows and pinned facts, clear old tool results, compact long runs,…
-
2
Long-Term Memory and Summaries
Help agents remember across sessions: running summaries, a facts store that updates instead of duplicating, extracting memories with an LLM, deciding what is worth remembering, and letting users…
-
3
Vector Memory with Embeddings
Give an agent a searchable memory: store memories as vectors, retrieve by meaning, and rank them by relevance, recency and importance, with a runnable scorer and a vector-database…