- This event has passed.
Reducing LLM Costs Through Programmatic Tooling
The Model Context Protocol (MCP) provides a standardized way for large language models (LLMs) to discover and invoke external tools through a client–server architecture. In its most common usage, MCP tools are called directly by the model during inference, one action at a time, via structured tool calls. While effective, this approach can become limiting when workflows grow more complex.
In this session, we’ll introduce MCP Code Mode, a paradigm shift in how LLMs interact with tools. Rather than emitting a series of discrete tool calls, the model generates executable code that invokes, sequences, and coordinates MCP tools directly. The result: clearer intent, tighter control flow, and more powerful orchestration.
We’ll start with a quick tour of MCP’s architecture and the standard tool-calling patterns most of us are familiar with, then dig into where those patterns begin to creak under real-world complexity. From there, we’ll explore how Code Mode unlocks a more expressive and efficient way for LLMs to reason about actions by using code as the glue.
To make it concrete, we’ll walk through live demos that compare direct tool calling with code-based orchestration, highlighting where Code Mode shines in practice and why it can be a game-changer for building robust, scalable AI systems.
About the Speaker
Eric Charles is an active contributor and committer to several open source projects, including Jupyter and Apache. He is the founder and CEO of Datalayer (https://datalayer.ai), a platform for AI-driven data analysis. Prior to founding Datalayer, Eric collaborated with leading SaaS companies to design and implement innovative open source solutions.
About the Open Source Architect Community
The Open Source Architect (OSA) Community is an invitation-only group for seasoned software architects who are passionate about open source technology. Request to join the OSA Community: https://forms.gle/7efbynVzYhhH2LCQ7
We review each application carefully. If it’s a fit, you’ll get an invite to join the space where it all happens.
For the latest updates on all things open source, follow our public feed on LinkedIn.