Toshi’s pick this week is Exploring LLMs in Chemical Engineering: Thermodynamic Modeling meets Vibe Coding from LinkedIn by Antonio Mazzei

Have you ever wondered what happens when you combine agentic AI workflows with MATLAB? A couple months ago, I saw a fantastic example from Antonio Mazzei that shows just how powerful this combination can be.

Antonio used a “vibe coding” approach with Gemini CLI to connect to the MATLAB MCP Core Server and build an interactive script for thermodynamic modeling. The goal was to visualize the 1-parameter Margules model for Excess Gibbs Free Energy—a classic chemical engineering problem.

What’s exciting here isn’t just the math—it’s the workflow. By letting an LLM act as an agent, Antonio didn’t just generate code; he created a system that understands context, iterates on solutions, and integrates with MATLAB. This is a glimpse into the future of technical computing.

The best part? The MATLAB MCP Core is open source and available on GitHub, so you can try this yourself here. Whether you’re in chemical engineering, finance, or any modeling-heavy domain, this approach opens up new possibilities for automation and creativity.

Your turn: Have you experimented with LLMs for modeling or simulation tasks? Share your experience in the comments—I’d love to hear how you’re using AI to push boundaries.