AI Needs Memory - Here's How It Works

Abstract // I'll explore the foundational decisions that determine whether AI agents deliver lasting value or become expensive technical debt. Rather than focusing on specific frameworks or tools, I'll cover the core tradeoffs between flexibility and performance, the memory patterns that actually matter for agent reliability, and how to architect systems that can evolve with the rapidly changing AI landscape. The key insight is understanding what problems agents fundamentally solve—automating complex, multi-step workflows—and designing memory and coordination systems around those core needs rather than getting caught up in today's specific technologies. You'll leave with a framework for making architectural decisions that will serve you well regardless of which models, frameworks, or tools become dominant next year. Bio // I'm a Principal Machine Learning Engineer at Workhelix, where I've been building enterprise-scale GenAI platforms and production ML systems as a founding engineer. I wrote the O'Reilly book "What Are AI Agents?" and am currently writing a follow-up called "Managing Memory for AI Agents" about the practical considerations of working with and managing data with AI agents. I recently published research in AEA Papers and Proceedings on measuring firm-level exposure to large language models and their potential productivity impacts. I spend most of my time figuring out how to actually deploy AI systems that work reliably in production—from async LLM APIs and embedding systems to the messy real-world challenges of putting GenAI into enterprise workflows. My work spans the full stack, and I split my interests between the theory of AI and its impacts on labor, to deploying production-grade LLMs and AI Agents An MLOps Community Production sponsored by Databricks