Overview In this role you will lead the architecture and feature definition for the out-of-band management of AMD Instinct accelerators and server platforms, shaping rack-scale management. You'll collaborate with cross-functional teams, partners, and customers to deliver novel manageability features and integration with datacenter infrastructure. You'll tackle end-to-end OOB challenges from DCIM software to device firmware, driving practical architectural outcomes. This is a high-visibility leadership position with impact on future AMD products and datacenter operation.
Responsibilities- Lead architecture and feature definition for OOB management of AMD Instinct GPUs and server platforms, including health, power/thermal, firmware lifecycle, inventory, and error handling.
- Define standards-based interface strategy for GPU/server manageability using Redfish, PLDM, MCTP, and related specs.
- Collaborate with BMC/firmware teams to define OOB requirements, including Redfish schemas, sensor/inventory representations, eventing, firmware update flows, and debug workflows.
- Contribute to firmware management architecture, covering in-band and out-of-band update flows, versioning, dependencies, activation, and recovery.
- Engage with customers, partners, and ODM/OEMs to understand datacenter integration needs and translate into concrete feature requirements.
- Work with architects, firmware/software, security/RAS, and validation teams to produce production-ready deliverables and contribute to conformance/validation strategy.
- Help shape AMD Instinct roadmaps through customer input, participate in standards/open-source communities, and mentor engineers and architects.
Key requirements- Expert-level experience in platform management architecture, server manageability, BMC or embedded firmware, or GPU/accelerator platform design.
- Strong understanding of the full OOB management stack from DCIM and orchestration to BMC firmware and device protocols.
- Deep knowledge of DMTF Redfish (schema design, OEM extension, eventing, update service) and PLDM/MCTP; OpenBMC experience preferred.
- Experience with firmware security concepts (secure boot, root of trust, firmware signing, attestation, SPDM) and architecture/validation delivery across teams and partners.
- Experience engaging with customers or OEM/partner teams to gather requirements and drive alignment; familiarity with AMD platforms, AI/HPC, or OCP rack architectures is a plus.
- Bachelor's or Master's degree in Computer Science, Computer Engineering, or Electrical Engineering (preferred).
- Excellent communication and presentation skills for leadership and customer interactions
- Ability to work cross-functionally and influence without formal authority
- Mentorship and capability-building across teams
- Redfish (schema design, OEM extensions, eventing)
- PLDM and MCTP for inventory, monitoring, control, and firmware updates
- OpenBMC architecture and services
