AI agents connect with development tools, documents, and workflows to help engineers design, debug, analyse, and deploy embedded systems.
AMD has introduced AMD Ross, an agentic artificial intelligence (AI) assistant for embedded development. It is designed to support engineers across system architecture, hardware design, debugging, software development, edge AI, and deployment. Unlike a coding assistant, Ross connects AI agents with development tools, engineering documentation, and workflows.
Engineers can interact with AMD development tools using natural language. Ross can search documentation, check tool status, run commands, assist with debugging, analyse results, and execute engineering workflows.
Ross supports AMD’s embedded portfolio, including field-programmable gate arrays (FPGAs), adaptive system-on-chips (SoCs), x86 embedded processors, and edge AI platforms. It can be used for hardware partitioning and design, software development, machine learning, power analysis, board design, and deployment.
The system uses Model Context Protocol (MCP) servers to connect AI agents with AMD Embedded development tools. MCP is an open standard that allows agents to access information, run commands, and interact with tools within a development environment.
Ross also includes an AMD Knowledge Base containing information from user guides, product guides, white papers, application notes, and answer records. The knowledge base can be accessed through cloud-based or locally hosted offline environments, allowing agents to use this information during development tasks.
Engineering teams can create agent skills for repeatable tasks. These Markdown files are written by experts to guide large language models through defined workflows. A skill, for example, can guide an agent through timing optimisation or restructuring a C++ design for higher performance in Vitis HLS. Teams can share these task-specific methods across projects.
Design examples show how these skills can be applied to embedded development and can serve as starting points for engineers building their own applications.
Ross can also work with developers’ preferred large language models, integrated development environments (IDEs), and command-line environments. It brings AI agents, AMD development tools, engineering knowledge, skills, and design examples into the embedded development process.
Click here for the original announcement.
