Embedded development spans hardware design, software, and debugging. AMD wants developers to manage more of that work using natural-language prompts.
The company introduced AMD Ross on Sept. 30, an agentic AI assistant designed to help engineers design, optimize, debug, and deploy embedded systems using supported AMD tools. According to AMD, Ross combines its knowledge base, expert-authored agent skills, design examples, and Model Context Protocol (MCP) servers that connect AI agents to development tools.
AMD says this combination can automate repetitive development tasks, freeing engineers to focus on more complex design work.
“AMD Ross brings AMD Embedded tools, trusted knowledge and expert-authored workflows together in a single agentic AI experience grounded in the technologies and methodologies our customers use every day. AMD Ross brings the power of agentic AI to embedded developers to move product innovations from design intent to deployment faster by accelerating the entire life cycle,” said Salil Raje, senior vice president and general manager, AMD Embedded.
AMD Ross is available now and requires no additional Ross license, although developers still need the applicable licenses for their AMD tools. It supports compatible AI clients, including Codex CLI, Claude Code, and GitHub Copilot CLI, and requires access to a language model.
How Ross works with embedded development tools
AMD Ross is designed to shorten development cycles through reusable agent skills, documentation access, and design examples, according to AMD. AMD aims to improve developer productivity by delegating standard workflows to AI agents.
Agent skills provide structured instructions for specific field-programmable gate array (FPGA) design tasks. Developers can extend existing skills or create their own to adapt workflows to their projects.
AMD’s knowledge base supports cloud and local access. For companies requiring air-gapped environments, AMD provides guidance for deploying the database and supporting components locally, with a suitable answer-generation model that can operate without an internet connection.
AMD lists support for all Vivado versions and Vitis HLS 2025.2 and later. Vivado workflows use an MCP server for live tool interaction, while HLS skills use standard command-line tools. AMD plans monthly releases to add tools and workflow capabilities.
What this means for the embedded systems space
As silicon vendors expand their developer tools, teams need to assess the practical benefits. For those using AMD tools, the question is whether Ross reduces time spent navigating documentation, running tool commands, and troubleshooting designs.
Teams can start with tasks such as design optimization and hardware debugging. Reusable workflows could also help engineers share knowledge across projects.
Before expanding adoption, test Ross on a supported workflow and compare the results with your existing process. Review generated code, recommendations, and proposed actions before applying them, and confirm that your model and deployment configuration are approved for the project data involved.
Read more: As teams evaluate tools like Ross, Microsoft’s AI coding-agent study highlights why regular use and review capacity matter when assessing productivity gains.