MCP: Bridging LLMs with Your Applications and Data - Java/Quarkus Version
summary
MCP is presented as a practical architecture for building context-aware systems that orchestrate LLMs for complex tasks. The talk explains how to design generic and domain-specific MCP servers, when to choose each, and how to structure prompts and interactions for best LLM performance, with concrete Java and JavaScript examples using LangChain4j and Quarkus and guidance on cloud deployment for scalability and reliability.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 6:34LLMs in development and privacy
- 10:32LLM evolution and tools
- 17:14What is MCP?
- 23:41MCP architecture
- 31:08MCP in applications
- 34:23MCP service types
- 37:14MCPs as traditional APIs
- 40:01Conclusion
related talks
39mCreate Intelligent MCP based applications with Quarkus
32m30 minutes to understand MCP (Model Context Protocol)
57mIntegrating LLMs in Java A Practical Guide to Model Context Protocol
02h 49mMCP Servers: Best Practices, Design Choices and Their Consequences
47mMCP Demystified: Cutting Through the Hype-What It Is, What Isn't & Its Role in the Agentic RAG Stack
12mMCP and the Dawn of Agentic Intelligence
32mMCP for dummies: Feed more context to LLM
46mMCP for LLM Agents, Infrastructure APIs & Data Services