In this episode of the AWS Developers Podcast, Sebastien Stormacq is joined by Giuseppe Battista and Kevin Shaffer-Morrison for a deep dive into Agentic AI and the Model Context Protocol (MCP). They explore how startups are rapidly adopting generative AI and the business and technical hurdles they encounter along the way. The conversation covers how AI agents are integrated into workflows, the role of orchestration technologies, and how tools enhance the planning capabilities of AI models. The trio unpacks the MCP framework and its potential to standardize communication between large language models and tools—drawing a compelling analogy to how USB transformed hardware interactions. They also discuss the complexities of tool orchestration, including resource management and security, and what the rise of ‘tool definition engineering’ means for the future. From integrating MCP with platforms like Amazon Bedrock to rethinking the role of software engineers, this episode looks at how AI is reshaping the foundations of modern software development.

With Kevin Shaffer-Morrison, Senior Startup Solution Architect ; With Giuseppe Battista, Senior Startup Solution Architect

Model Context Protocol (MCP) and Amazon Bedrock

Running MCP-Based Agents (Clients & Servers) on AWS

The multi-agent orchestrator

The podcast where we talk about the multi-agent orcuestrator

Agent Orchestration: CrewAI

Agent Orchestration: Pydantic

Smolagents

Swami's post on Linked about MCP