Meetup: Memory, Context & Agents (SurrealDB x CCCL)

Building an AI agent is the easy bit. What trips people up is everything that comes after — how do you keep context across sessions without blowing your token budget? How do you reason over relationships without a graph database bolted onto the side? How do you build something that actually behaves like it has memory? Claude Code Bengaluru and SurrealDB are coming together to get into the weeds on that. Two talks from engineers who've been building real agent systems on SurrealDB — one demoing a persistent memory engine for Claude agents, the other building a multi-model CRM agent from scratch. Plus two more talks from the CCCL community. Talks Farhan Khan (SurrealDB) — SurrealDB as the context layer for AI agents Abhishek Das (SurrealDB Community, Bengaluru) — Enabling Multi-Model Agents on Mastra with SurrealDB CCCL Speaker — TBC Agenda Doors - 9:30 AMTalks - 10:30 AM Food & networking - 12:00 PMClose - 1:30 PM About the speakers Farhan Khan (SurrealDB) Farhan is on the core engineernig team at SurrealDB and has been building a persistent memory engine for Claude agents on top of SurrealDB — hybrid search, knowledge graph, lifecycle-aware context, all in a single database. He's been using it in his own workflow and will show you exactly how it works. Abhishek Das (SurrealDB Community, Bengaluru) Abhishek is SurrealDB's community lead in Bengaluru and has been deep in agentic frameworks. His talk builds a working CRM agent from first principles using Mastra and SurrealDB — graph traversal for relationship-aware lookups, no separate vector store, no brittle ETL pipelines. Live demo: an agent that answers "who should I call today, and why?"