Agent Memory Hack Night | Mastra x Elasticsearch
- WHO IT IS FOR: AI developers, TypeScript engineers, and LLM architects looking to improve agent persistence and context.
- WHAT IT IS ABOUT: Integrating Mastra and Elasticsearch to build AI agents with advanced long-term and short-term memory capabilities.
- FORMAT: A hands-on hack night featuring a brief introduction and demo followed by an open building session.
Stop building AI agents that start every conversation from scratch. Join Mastra and Elasticsearch in San Francisco for a focused hack night dedicated to solving the challenge of agent memory. You will learn how to move beyond stateless interactions and build agents that remember user preferences, track progress toward goals, and recall previous decisions to make smarter, context-aware choices in the present.
Leveraging the Mastra open-source TypeScript framework and the powerful retrieval capabilities of Elasticsearch, you will explore how to implement native vector stores and hybrid search. The session will dive into the nuance of memory decay and relevance, ensuring your agent prioritizes recent information over outdated data for more accurate results.
Whether you are building a personalized learning coach, a sophisticated customer support assistant, or a research agent that tracks complex findings, this event provides the tools and environment to bring your project to life. Bring your own data or use a starter kit to start building immediately after a quick introductory demo.