• Speaker: Justin Castilla
  * Room: CC 234
  * Time: Sat 2:00 pm – 2:30 pm

• Format: Lecture (30 Min + Q&A)
  * Difficulty: Introductory
  * Track: AI / ML
  * Presenter Location: In-person
  * Experience: umpteenth time speaking

Modern AI applications are evolving beyond single chatbots into networks of specialized agents working together. Agent2Agent (A2A) is a protocol that enables AI agents to communicate and collaborate — for example, in a digital newsroom where a Reporter agent writes stories, a Researcher gathers facts, an Editor reviews quality, and an Archive agent retrieves historical context. A2A coordinates who talks to whom.

But each agent also needs access to tools and data. This is where Model Context Protocol (MCP) comes in, providing standardized access to databases, APIs, and other resources. MCP handles what tools each agent can use.

This talk demonstrates how A2A and MCP complement each other through a practical newsroom example. Attendees will learn:

• When to use A2A for multi‑agent coordination
  * When to use MCP for tool and data integration
  * How to implement both protocols in real applications
  * Why combining them leads to more maintainable, scalable systems

Developers will leave with a clear understanding of how to build distributed agent systems that leverage specialization, autonomy, and standardized tool access to accomplish complex workflows that single agents struggle to handle efficiently.

Target Audience:

• AI/ML developers
  * ML DevOps practitioners