Upcoming in Boston

Crashproof Agents and Agentic Memory

An evening for people building, shipping, and scaling with AI. Two talks on what it takes to keep agents running once they hold real state. Alex Poliakov (DBOS) on Durable Execution: what it guarantees, what it does not, and how DBOS achieves durability using only Postgres. Alex Seriy (Cockroach Labs) on the data tier underneath agent memory: surviving node, zone, and region failures, scaling by adding nodes, and reading data exactly as it was hours ago. Supported by HackerSquad and Builders Collective.

RSVP on Luma See agenda
Presenters

Who shared what

AP
Alex Poliakov
Head of Customer Solutions · DBOS Inc.

Head of Customer Solutions at DBOS, working with teams that build durable, long-running AI agents on Postgres.

Talk

Building Crashproof Agents

Beyond the problems of accuracy and cost, AI agents are challenged by long-running execution processes, human-in-the-loop approval logic, non-deterministic call paths, and upgrade complexity. Agents are expected to perform more important tasks over time, with increasing costs of failure. These challenges can be met with Durable Execution, a set of techniques that originated in transactional processing and is now growing in popularity with agent builders. This talk introduces durable execution and the guarantees it can, and cannot, provide, then shows why durability is becoming a key pattern in AI agents. It closes with how DBOS achieves durability using only Postgres, including advanced maneuvers like queueing, upgrading long-running workflows, and forking. For developers with experience in agentic development or an interest in reliable software. Postgres experience is a plus.

Speaker profile →
AS
Alex Seriy
Senior Staff Sales Engineer · Cockroach Labs

Senior Staff Sales Engineer at Cockroach Labs, where he helps organizations use CockroachDB for scalability, resilience, and performance.

Talk

Agentic Memory: The Final Frontier

Once agent and workflow state live in your database, three questions become sharper: what happens when a database server goes down, what happens when the fleet grows 100x, and how do you find out what your agent was actually looking at three hours ago. This talk is about CockroachDB, a distributed SQL database that speaks Postgres, and what the data tier looks like when it has the same properties we spent the last decade building into everything above it. It covers how it survives node, zone, and region failures with no manual intervention and no data loss, how it scales by adding nodes instead of sharding or re-architecting, and how it lets you read data exactly as it was in the past, which is useful for both debugging non-deterministic systems and answering auditors. It closes on what it means to keep transactional data, workflow state, and vector embeddings in one place instead of three.

Speaker profile →
Agenda
  1. 6:00 PM Networking and food
  2. 6:30 PM Welcome, Kranthi Manchikanti
  3. 6:35 PM Talk 1 · Alex Poliakov, Building Crashproof Agents
  4. 7:15 PM Talk 2 · Alex Seriy, Agentic Memory: The Final Frontier
  5. 7:55 PM Open discussion · networking
  6. 8:30 PM Close

RSVP on Luma Add to calendar

Platform partner
DBOS Featured speaker Cockroach Labs Featured speaker HackerSquad Supported by Builders Collective Supported by
Subscribe

Get the recap

Slides, recording, and the recap will land here within a week, and in subscribers' inboxes first.