TWed Talk: "Model Context Protocol (MCP): Standardizing Tool Use for LLM Systems" (4p, Weds, 18 Feb 2026)

WHAT: TWed Talk: "Model Context Protocol (MCP): Standardizing Tool Use for LLM Systems"

WHO: Henrique Santos

WHEN: 4p, Weds, 18 Feb 2026

WHERE: Winslow 1140

WEBEX: https://rensselaer.webex.com/meet/erickj4

EVENT PAGE: https://bit.ly/4tvCDAt

 

Please join us Weds, 18 Feb (4p) as Henrique Santos leads us in a hands-on introduction and demo of the Model Context Protocol (MCP), an impressive new open-source standard for connecting AI applications to external systems including data sources, tools, and workflows. Pizza arrives approx. 3:30p, the talk begins at 4p. 

 

DESCRIPTION: Large language models are increasingly integrated with tools, databases, and external services. However, tool integration is often ad hoc and model-specific. The Model Context Protocol (MCP) proposes a standardized interface for exposing tools, resources, and prompt templates to LLM-based systems. This hands-on talk introduces the motivation behind MCP, explains its core abstractions (tools, resources, prompts), and explores how modern runtimes (e.g., LM Studio) orchestrate model–tool interaction. We conclude with a live demo showing MCP integration in practice.

 

BIO: Dr. Henrique Santos is Director of Semantic Applications Research at Rensselaer Polytechnic Institute. His research focuses on semantic web technologies and knowledge graphs for artificial intelligence, data interoperability, and scientific data integration. He has led work on projects funded by DARPA, IARPA, and NIH, including Machine Common Sense, ECOLE, HIATUS, and RCADS. His recent work explores how large language models can integrate with semantic infrastructures to improve interpretability and reproducibility in research.