AMD unveils Ross agentic AI for embedded engineering

AMD has introduced AMD Ross, an agentic AI assistant designed to streamline embedded systems engineering. The tool acts as an abstraction layer spanning hardware, software, and system-level engineering to assist design engineers working with AMD Field Programmable Gate Arrays (FPGAs), adaptive SoCs, embedded x86 processors, and edge AI platforms.

Unlike general-purpose AI copilots, Ross is tightly integrated with AMD’s specific toolchain, including Vivado, Vitis, ROCm, and Ryzen AI software. It utilizes Model Context Protocol (MCP) servers to allow AI agents to interact directly with design environments, enabling them to run commands, analyze reports, and orchestrate complex workflows.

The assistant is intended to support a wide range of tasks beyond simple code generation, including hardware debugging, signal capture, power estimation, and machine-learning algorithm optimization. It can assist with analyzing timing violations, recommending architectural refinements in Vitis HLS, and providing guidance on schematic reviews and board layouts. AMD Ross is available through both cloud and local offline environments.