AMD Launches ‘Ross’ Agentic AI Assistant to Accelerate Embedded Design and Development
AMD has introduced AMD Ross, a purpose-built agentic AI assistant designed to dramatically accelerate the entire embedded systems development lifecycle—from early architectural design intent to final field deployment.
Spanning the company’s embedded hardware portfolio—including FPGAs, adaptive SoCs, x86 embedded processors, and dedicated edge AI platforms—AMD Ross enables engineers to interact with complex toolchains using natural language. Rather than serving merely as a generic code copilot, the platform connects directly into AMD development environments to automate repeatable workflows, run tool-level commands, and perform deep root-cause troubleshooting.
“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, 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. It 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.”
The architecture of AMD Ross relies on four foundational components:
- Model Context Protocol (MCP) Servers: Built on the open-standard MCP framework, these servers connect autonomous AI agents directly to native AMD Embedded tool suites, allowing the system to query status, execute commands, and iterate within active design environments.
- AMD Knowledge Base: A validated, vectorised repository of user guides, datasheets, application notes, and answer records that delivers verified technical guidance either via cloud connectivity or entirely offline in secure, on-premise environments.
- Agent Skills: Reusable, expert-authored markdown instructions capturing institutional engineering best practices for complex design challenges—such as timing closure and C++ refactoring for Vitis HLS.
- Ready-to-Run Design Examples: Pre-built reference architectures demonstrating practical agentic workflows across real-world edge compute and embedded deployments.
To ensure seamless integration within diverse development stacks, AMD Ross is client-agnostic. Design teams retain the flexibility to deploy their preferred Large Language Models (LLMs), integrated development environments (IDEs), and command-line interfaces while maintaining full operational hooks into AMD’s toolsets.
Beyond software authoring, AMD Ross assists engineering teams with hardware/software partitioning, schematic reviews, board layout optimization, low-power estimation, and high-level synthesis (HLS) hardware implementations. In security and performance-critical deployments, it accelerates debugging by identifying timing violations, configuring debug cores, and offering contextual fixes.
“The AMD Ross agentic AI assistant has been a valuable addition to our development workflow,” said Geetha Govindaraj, Associate Director, FPGA SOM BU at iWave Global. “It helps us identify and troubleshoot issues more efficiently, reducing the time and effort required during the debugging and bring-up stages.”
AMD Ross is available immediately for system developers and engineering teams, with AMD planning to roll out expanded embedded tool integrations and workflow capabilities on a monthly release cadence.
