Now that the trading agent is deployed, let’s give it a memory. On Agent Platform you get session persistence for free — ask it “what trades did you execute?” a day later from the same session and the whole conversation is still there, even though the agent was suspended the entire time (so you’re not paying for it to sit idle). From there we’ll add proper cross-session memory with ADK’s callbacks, then make it deterministic: leaning on a model to decide what’s worth keeping is a little hit-and-miss, and for a trading agent the trades are exactly the thing we want to remember every single time. Here’s how.
Scroll horizontally to explore. Open full-size diagram ADK Agents will automatically have session persistence when deployed on Agent Platform. You can test this by asking follow up questions in the running sessions like “What trades did you execute?”. The full conversation … Continue reading Agent memories with ADK
In the previous post we sent the agent’s session events to Memory Bank with add_events_to_memory() in an after-agent callback. An LLM extracts useful facts from those events, and embeddings make the saved facts searchable in a later session. That gives us general recall, but the … Continue reading Deterministic memories