Détails de la formation
This course introduces how agentic AI can be integrated into MATLAB-based design workflows to accelerate analysis, implementation, and verification while keeping engineers in the loop. The concepts and techniques are designed to be transferable across large language models and AI tools, enabling participants to apply what they learn with the technologies that best fit their organization.
Participants will learn:
- How MATLAB ® can be effectively used by AI agents through Model Context Protocol (MCP)
- How planning, execution, testing, and iteration enable reliable agentic engineering workflows
- How to use agents as collaborators for solving realistic engineering problem
- How to customize agentic experience with engineering knowledge using skills
Day 1 of 1
Introduction to Agentic AI
Objective: Introduce the fundamentals of agentic AI and the key concepts needed to perform workflows in MATLAB.
- Chatbots vs. agentic AI systems
- The agentic loop: plan, generate, and evaluate
- Context in agentic AI
- Skills and Model Context Protocol (MCP)
- Overview of MathWorks agentic capabilities
Connecting MATLAB with AI Agents
Objective: Set up the MATLAB Agentic Toolkit and use it to complete tasks in MATLAB.
- MATLAB Agentic Toolkit installation
- MATLAB MCP Core Server and available MATLAB skills
- MCP interaction basics and key considerations
- Getting started example
Solving Engineering Problems in MATLAB with Agentic Workflows
Objective: Use agentic AI to solve real engineering problems in MATLAB with a human-in-the-loop approach.
- Project setup, scope, and source control
- Framing engineering problems for agentic workflows
- Formulating and executing a plan to solve an engineering problem
- Evaluating, testing, and documenting agent-generated MATLAB code
Skills, Context, and Customization
Objective: Create custom skills for repeatable workflows and understand how to scope them effectively.
- Overview of skills: files, structures, and context
- Authoring custom skills
- Defining structured workflows for repeatable behaviors
- Embedding domain expertise in skill design
Niveau: Intermédiaire
Pré-requis:
Fundamental knowledge of MATLAB
Durée: 1 jour
Langues: English, 日本語
