Beyond the Context Window: Building Shared Agent Memory

Your coding agent learns how a project works, then the next session repeats the same investigation. Another agent misses an architectural decision, follows an outdated instruction or cannot access knowledge held in a different tool. Keeping more conversation history does not solve all of these problems—and sharing everything introduces problems of its own.

In this three-hour hands-on workshop, you’ll build a shared memory layer for agents working on a software project. Starting with a supplied repository and agent harness, you’ll distinguish conversation history, agent instructions and durable project knowledge, then connect two agents to a common knowledge store. You’ll explore how an open-source approach lets you own, inspect and move that knowledge while keeping it usable across different agent tools.

You’ll work through the memory lifecycle: deciding what to retain, retrieving relevant information, recording its source, updating outdated knowledge and controlling which agents can read or change it. You’ll introduce conflicting guidance and private engineering notes, inspect what each agent retrieves, and add access controls and a review step before proposed findings become shared knowledge.

Finally, you’ll repeat tasks in fresh sessions and compare results with and without shared memory. You’ll measure task correctness and context size, examine retrieval traces, and investigate failures caused by missing, stale or inaccessible information. We’ll use OpenLore as an open-source example, with the emphasis on design principles and trade-offs you can apply to other memory systems.

You’ll leave with a working repository, a shared memory setup, repeatable evaluation fixtures and a practical method for deciding what your agents should remember—and whether that memory actually helps.

Who it’s for

Software and AI engineers building coding assistants or applications with multiple agents. You should be comfortable reading TypeScript and working in a terminal. Experience using an AI coding agent is helpful; no previous experience with agent memory systems is required.

What you’ll need

Bring a laptop with Node.js, Git and a local development environment. You’ll need access to a coding agent and a model API key for the supplied harness. Installation instructions and sample project knowledge will be provided before the workshop.

About the presenter

Shivay Lamba

Shivay Lamba is a software engineer and open source contributor passionate about AI and edge computing. With experience across startups and enterprise tech, he focuses on simplifying complex technologies for developers. Shivay actively speaks at global conferences, organizes community events, and contributes to projects in cloud-native ecosystems, WebAssembly, and machine learning. He has also built and mentored educational programs to bridge gaps in emerging tech adoption. When not coding, he enjoys writing, traveling, and mentoring young technologists.