Staff Software Engineer at LinkedIn, Seattle, Washington, USA

Title of the Talk:
Agentic AI: From Foundations to Practical Application – Building Intelligent Agents for Modern Workflows

Abstract of the Talk:
This keynote introduces participants to the emerging field of agentic artificial intelligence, which extends traditional AI by incorporating goal-oriented behavior, autonomy, and adaptability. The session begins with an exploration of the core capabilities of agentic AI systems, perception, reasoning, planning, memory, and action, while contrasting them with conventional model-centric approaches. Attendees will delve into the architectural foundations of both single-agent and multi-agent systems, with particular emphasis on how open protocols such as the Model Context Protocol (MCP) and Agent2Agent (A2A) allow agents to access external data sources and interoperate across platforms. The session also surveys leading frameworks like LangChain, LangGraph, AutoGen, and CrewAI, evaluating their features, memory integration, and support for human-in-the-loop workflows.

.