Spring Office Hours S5E04 - Spring AI Agentic Patterns with Christian Tzolov
10 snips
Feb 3, 2026 Christian Tzolov, a Spring AI engineer working on the MCP Java SDK and AI tooling, discusses agentic patterns for building autonomous AI that can reason, plan, and execute tasks. He covers agent architectures, skills and progressive disclosure, tool calling and ACP integration, vector stores versus tool use, and best practices for authoring and sharing reusable skills.
AI Snips
Chapters
Books
Transcript
Episode notes
Spring AI Prioritizes Portability And Extensibility
- Spring AI focuses on portable, opinionated abstractions that work across models and vendors.
- Christian emphasizes extensibility so Java developers can use AI without leaving the Java ecosystem.
Use Local Agent Tools With Your IDE
- Use multiple models and local tools for daily developer tasks like reviews and code suggestions.
- Pair small agentic tools with your IDE to keep fast feedback loops and control costs.
Tool Calling Shifts Context Curation
- The key AI challenge is curating context for a stateless, text-in/text-out foundation model.
- Tool calling shifts context curation from humans to models by letting models request needed data dynamically.
