Agents in Production

Agents in Production: Durable State

​Nothing in a demo has to survive the night. Production agents run for days, get restarted, get deployed over, and are still expected to pick up exactly where they left off. ​For AI Engineer week, SurrealDB is teaming up with AWS and partners on the question underneath all of it: where does an agent's state actually live once the process that created it is gone? ​Talks from the people building the runtime, the framework, and the layer underneath. Partners and talks are announced soon, so treat this as your placeholder for the 12th. ​Expect honest talks, live demos, plus food, drinks, and plenty of time for networking.

Agents in Production: Shared Memory

Your agent is right in the demo. Then it hits production, a fact changes underneath it, and it answers confidently and wrong. For SF Tech Week, SurrealDB is teaming up with Mastra on the seam between the two layers this actually lives in. Mastra runs the agents and decides what each one needs to know. SurrealDB holds what they know and keeps it true. Get that boundary wrong and no amount of prompting saves you. Expect honest talks, live demos, plus food, drinks, and plenty of time for networking.

Agents in Production: Trust, Access & Control

Agents in Production: The Full Stack

Agents in Production: Scaling Enterprise AI

Agents in Production: The Memory Problem @ NY Tech Week

Agents in Production: The Memory Problem

Agent memory is a mess. Most teams are stitching together a vector store, a graph database, and some middleware - three systems to manage, three potential failure points, and no guarantees that any of it stays consistent. It works until it doesn't. In this meetup, we're looking at what it actually takes to solve this properly. Not a quick fix, not another abstraction layer - memory that's transactional, structured, and built into the database itself. We'll look show what that looks like in practice with Spectron, SurrealDB's brand new agent memory system. We'll also walk through a full-stack knowledge graph RAG system - end-to-end, real code, real trade-offs - plus a guest talk from LangChain. Hard questions welcome. Drinks & pizza on us. Talks Tobie Morgan Hitchcock (CEO, SurrealDB) Spectron: What agent memory looks like when it's built into the database Martin (Solutions Engineer, SurrealDB) Full-stack knowledge graph RAG with agentic memory Igor Repka (Solutions Architect, LangChain) Details coming soon! Agenda 18:00 - Doors + drinks + pizza 18:30 - Welcome 18:35 - Talks + Q&A 20:00 - Drinks + networking 20:30 - Close About the speakers Tobie Morgan Hitchcock (CEO & Co-Founder, SurrealDB) Tobie Morgan Hitchcock is CEO & Co-Founder of SurrealDB, an AI-native, multi-model database for modern applications. A tech entrepreneur and software engineer with 17 years in the software and cloud-computing industries, he founded SurrealDB in 2021 with a focus on distributed databases and highly-available architectures. Martin Schaer (Solutions Engineer, SurrealDB) Martin is a computer science engineer working at SurrealDB and his own GenAI startup. He recently worked in lab automation designing declarative frameworks for instrument drivers, and his background spans everything from 3D visualisation for robotic systems to founding an advertising agency in Costa Rica.

Agents in Production: correctness, context, and control

Getting an agent to produce output is the easy part. Keeping it correct as data changes, memory grows, and systems scale is where things get interesting. This meetup is about what actually holds up in production: fresh context, durable memory, validation, observability, and guardrails that don't collapse under real-world pressure.

Agents in Production: workflows, memory & models

We've all done a lot of talking in the last few months (years??) about coding agents, advancements in models, and AI slop. In this meetup, we're bringing all of this together as our speakers talk about how to squeeze even more out of AI so you can actually ship something you're proud of.

Knowledge Graphs & Agents

We're kicking off our first meetup of 2026 with a practical deep dive into knowledge graphs + agents. The focus will be on clean graph retrieval patterns: querying a knowledge graph without unnecessary abstraction, keeping type-safety, and getting end-to-end observability (not just LLM calls).