Por que um agente de IA sem memória é uma calculadora cara. Construindo memória persistente

Vladislav Yankovsky, product tech lead of Evolution AI Agents at Cloud.ru, talks about building persistent memory for AI agents. He notes that memory is not needed by itself, but as a consequence of functional requirements: if an agent can work without memory between sessions, then it is not needed. He gives the example of an agent for Grafana, which does not need memory, and an L2 support agent, for which it is critical. He distinguishes between context (within the model window) and memory (outside the model), as well as two dimensions of memory: lifespan (short-term, working, long-term) and content type (semantic, episodic, procedural). In 80% of cases, a combination of short-term and long-term semantic memory is sufficient. He describes memory levels: level 0 — simple storage of dialogue history, insufficient due to duplicates and lack of consolidation; level 1 — vector search using pgvector, Qdrant, Milvus, Pinecone; level 2 — structured memory with SQL and graphs, for example, Graphiti. The key mistake is to store raw responses; the right approach is to extract facts and maintain a two-layer model with sources and extracted data.