United States – AMD has introduced AMD Ross, an agentic AI assistant designed to support embedded system development across hardware, software, AI and system-level design.
The assistant is intended to support development activities ranging from architecture and hardware design to optimisation, debugging, PCB design, schematic review, software development, AI implementation and system deployment.
AMD Ross uses natural-language interactions to connect developers with AMD Embedded tools, documentation and workflows. It also supports expert-developed agent skills, reusable workflows and an AMD Knowledge Base designed to provide contextual guidance throughout the development process.
“Embedded development is becoming increasingly complex as teams work across hardware design and debug, software development and deployment, AI inference, and system-level design,” said Salil Raje, Senior Vice President and General Manager at AMD Embedded.
“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,” Raje continued.
Supporting embedded development workflows
AMD Ross supports development across AMD’s embedded portfolio, which includes FPGAs, adaptive SoCs, x86 embedded processors and edge AI platforms.
Through natural-language commands, developers can use AI agents to search documentation, check tool status, execute commands, assist with debugging and run established workflows.
The platform supports tasks including hardware and software partitioning, high-level synthesis-based hardware development, silicon design and debugging, embedded software and algorithm development, machine learning optimisation and power optimisation.
It also supports system schematic review and board layout optimisation.
AMD Ross is designed to be client-agnostic, allowing development teams to use their preferred large language models, integrated development environments and command-line environments while connecting to AMD Embedded development tools.
AMD-specific knowledge and workflows
AMD said Ross differs from generic AI assistants through its integration with AMD development environments, documentation, engineering knowledge and expert-authored methodologies.
The platform brings together four components: Model Context Protocol (MCP) servers, the AMD Knowledge Base, agent skills and design examples.
MCP servers connect AI agents with AMD Embedded tools, allowing the system to query information and execute commands within development environments.
The AMD Knowledge Base consists of AMD-validated databases containing user guides, product guides, white papers, application notes and answer records. The knowledge can be accessed through the cloud or in an offline local environment.
Agent skills are expert-authored Markdown files that capture repeatable engineering practices for tasks such as timing optimisation and restructuring C++ designs for performance in Vitis HLS.
Design examples provide ready-to-run implementations showing how these agent skills can be applied to embedded applications.
Development and debugging use cases
AMD said Ross is designed to support faster prototyping, shorter debugging and optimisation cycles, developer productivity and the onboarding of engineers by making established workflows and engineering knowledge more accessible.
The company also said the platform can help teams apply consistent practices across projects and reuse institutional engineering knowledge.
“The AMD Ross agentic AI assistant has been a valuable addition to our development workflow. It helps us identify and troubleshoot issues more efficiently, reducing the time and effort required during the debugging and bring-up stages,” said Geetha Govindaraj, Associate Director, FPGA SOM BU at iWave Global.
Example workflows include translating natural-language requirements into development actions, searching documentation, executing tool commands, generating code and iterating designs.
Ross can also help developers interpret tool errors, analyse timing issues, optimise Vitis HLS designs and support hardware debugging through guided signal capture, debug-core configuration and interface analysis.
AMD Ross is available now, with developers able to access the platform through AMD’s website. AMD said it plans to add further AMD Embedded tools and workflow capabilities to Ross on a monthly basis.
The company positions the platform as a way for development teams to apply agentic AI across hardware, software, AI and system design while making engineering knowledge and established workflows easier to access and reuse.
