Getting Started with GitHub Copilot for RF EDA

In this introductory tutorial, you'll learn how to use GitHub Copilot to automate RF design and simulation workflows using Keysight ADS (Advanced Design System) 2027 and Keysight Nexus 2027. The video demonstrates how AI-powered agents can interact directly with Electronic Design Automation (EDA) software to create schematics, configure simulations, access documentation, run analyses, and generate results with minimal manual effort.

The tutorial begins with an overview of the software and tools used in the demonstration, including the GitHub Copilot desktop application and the latest releases of ADS and Nexus. Special attention is given to the new Model Context Protocol (MCP) servers introduced in the 2027 versions of ADS and Nexus. MCP provides a standardized interface that enables AI agents such as GitHub Copilot to communicate directly with EDA tools, execute commands, retrieve documentation, and perform engineering tasks.

You'll also get a guided tour of the GitHub Copilot desktop application, including the main chat interface, context management features, interactive and autopilot modes, planning capabilities, model selection options, reasoning controls, repository connections, automations, customizations, and account settings. These features help users understand how to tailor Copilot's behavior for engineering workflows and productivity tasks.

A major focus of the video is configuring ADS and Nexus MCP servers within GitHub Copilot. Rather than manually editing configuration files, the demonstration shows how engineers can leverage Copilot itself to perform the setup process, including selecting specific ADS and Nexus installations when multiple versions are present on the system. After configuration and validation, the MCP servers become available to GitHub Copilot, allowing seamless communication with both simulation environments.

The second half of the video showcases practical RF design automation. Using an existing ADS workspace containing a transistor design, GitHub Copilot is instructed to create and execute a DC I-V sweep simulation. Viewers can observe the agent accessing product documentation, generating simulation configurations, creating schematics, executing simulations, and using Python and Matplotlib to generate plots and summarize results. The demonstration highlights best practices for working with AI agents in EDA environments, including strategies for prompting and simulation setup.

The workflow is then repeated using Nexus to illustrate how the same AI-driven approach can be applied across multiple simulation platforms. GitHub Copilot automatically creates a Nexus project, configures simulation tasks, runs the analysis, and compares the results against those generated in ADS. The comparison confirms that both tools produce virtually identical outcomes, demonstrating the interoperability and effectiveness of MCP-enabled automation.

By the end of the video, viewers will understand how to connect GitHub Copilot to ADS and Nexus, configure MCP servers, automate simulation workflows, and leverage AI agents to accelerate RF design tasks. Whether you're an RF engineer, microwave designer, simulation specialist, or EDA user looking to improve productivity, this tutorial provides a practical introduction to the future of AI-assisted design automation.