AI Dev 25 | Apoorva Joshi: Building Agents That Learn—Managing Memory in AI Agents

AI agents are redefining what’s possible with LLM-powered applications—moving beyond automation toward systems that can adapt, personalize, and learn over time. In this talk, Apoorva Joshi, Senior AI Developer Advocate at MongoDB, explores the central role of memory in enabling these intelligent behaviors. She breaks down the differences between short-term and long-term memory in agentic systems, and shares practical strategies for building agents that can retain knowledge, learn from their past, and evolve with each interaction. This session is ideal for developers interested in building more capable, context-aware AI agents—and understanding the memory mechanisms that power them. Keep building, DeepLearning.AI