​2-Hour Build: AI Agent Memory

​We’re hacking on agent memory this week.

​LangGraph is powerful, but agents forget everything between runs unless you wire up memory. In this session, we’ll give your agent a reliable “brain” that survives restarts, rolls back mistakes, and branches into parallel strategies. All built live, in the same room.

​→ Who’s this for?

​Anyone building with LangGraph (or planning to) who’s run into the “why does my agent forget?” problem and wants hands-on patterns to fix it.

​In these 2 hours, you will:

• ​Understand the core memory patterns behind stateful LangGraph agents.

• ​Learn by building: we’ll wire up a checkpointer, branch three what-ifs, and merge results.

• ​Leave with a clean starter repo you can drop into your own workflow.

​→ What we’ll hack on (together)

• ​Persist – plug in a checkpointer and make memory stick.

• ​Rewind & diff – trace an agent’s reasoning, tweak, and replay.

• ​Branch & merge – test multiple paths, keep the best one.

• ​Hot-migrate – evolve schemas without downtime.

• ​Compact – keep storage tidy and costs low.

​→ What to bring

​Your laptop. This is hands-on. We’ll demo with Convo SDK as the checkpointer, but the concepts are tool-agnostic.

​→ Host

​Anil Dukkipatty: Has worked with 800+ developers on agent memory and LangGraph workflows.

​→ Venue Partner

​
B-Hive - powering conversations that build the future of AI.

​→ Built by Builders, Backed by the Best

​​We’ve hosted engineers, founders, and researchers from some of the world’s most forward-thinking companies. Here’s a look at who’s been in the room.