Decentralized Data for a Centralized Brain with Valliappa Lakshmanan скачать в хорошем качестве

Decentralized Data for a Centralized Brain with Valliappa Lakshmanan

Watch the entire Superstream: https://learning.oreilly.com/videos/d... Agentic AI is forcing data engineers to rethink everything—from how data is modeled and stored to how it’s delivered and consumed. In this keynote from the Data Superstream, Valliappa Lakshmanan, CTO of ObinAI, explains why building data infrastructure for large language models and AI agents requires a complete mindset shift. Instead of optimizing for software engineers and data scientists, today’s systems must deliver in-context, semantically rich data directly to AI agents and business users. Valliappa unpacks how growing context windows, tool calling, and few-shot learning are replacing traditional analytics and ETL pipelines. He highlights why normalized tables and strict taxonomies may actually hinder model performance, and how to build example stores, vector indexes, context curation workflows, and clean APIs that align with the strengths and weaknesses of AI agents. If you’re responsible for making data usable inside agentic applications, this talk is your guide to building a modern data stack that supports perception and action—not just storage and retrieval. Follow O'Reilly on: LinkedIn:   / oreilly   Facebook:   / oreilly   Instagram:   / oreillymedia   BlueSky: https://bsky.app/profile/oreilly.bsky...