New The Skills of Tomorrow: how AI-exposed is every skill in 2026? See the data →
NVIDIA

Senior Platform Software Engineer - Platform Architect

NVIDIA
Apply →
remote senior full-time Bengaluru

First indexed 19 Aug 2026

Description

NVIDIA is building a new wave of accelerated computing systems where platform manageability, secure firmware, and high-performance networking must work as one coherent architecture.

We are looking for a Senior Platform Software Engineer - Platform Architect to help define and implement OpenBMC-based management for systems that include host CPUs, BMCs, and coordinated top-of-rack Ethernet switching functionality.

Responsibilities

  • Architect and implement OpenBMC support for platforms where the BMC handles host system functions and a coordinated ToR Ethernet switch subsystem.
  • Define the end-to-end switch-management architecture across OpenBMC D-Bus services, bmcweb/Redfish APIs, persistent configuration, eventing, telemetry, security, and backend switch control interfaces.
  • Develop OpenBMC object models and services for switch inventory, switch ASICs, front-panel ports, internal ports, PHYs, optics, retimers, VLANs, LAGs, FDB entries, counters, sensors, and health state.
  • Implement L2 networking capabilities including port admin control, speed and FEC configuration, auto-negotiation, breakout, VLAN creation and membership, access/trunk mode, MAC learning, FDB management, flooding controls, link aggregation, and loop-prevention features.
  • Build backend abstraction layers that can integrate with Linux switchdev/netlink, SAI/SONiC/SWSS-style components, DENT-style Linux networking components, or vendor switch ASIC SDKs.
  • Work closely with hardware, ASIC, security, firmware, QA, manufacturing, and cloud infrastructure networking teams to influence board architecture, schematics, manageability interfaces, and production test strategy.
  • Write and review design specifications, API documentation, code, unit tests, hardware-in-loop tests, integration tests, and QA plans required to productize the architecture.

Requirements

  • Bachelor's degree or higher, or equivalent experience, in Computer Science, Electrical and Electronics Engineering, or a related field with 12+ years of relevant experience.
  • Strong OpenBMC firmware development experience on x86 or ARM-based server environments, such as Yocto, systemd, D-Bus, phosphor services, bmcweb, Redfish, firmware update, device monitoring, and platform security.
  • Hands-on experience with Ethernet switch architecture and L2 features such as VLANs, trunks, access ports, LAG/LACP, STP/RSTP, LLDP, FDB/MAC learning, storm control, ACLs, QoS, PHY management, and transceiver diagnostics.
  • System management software experience with rack or blade server platforms, including IPMI, KCS, Redfish, PLDM, MCTP, SPDM, secure firmware update, attestation, and BMC-host communication.
  • Strong programming and debugging skills in C/C++ and Python, with familiarity in Bash, Go, or Rust; ability to develop reliable Linux user-space services and review low-level system code.
  • Excellent written and verbal communication skills, strong ownership, high engineering judgment, and a passion for solving hard multi-functional problems.

Nice to Have

  • Contributions to OpenBMC, Linux networking, switchdev, SONiC, SAI, DENT, FRR, LLDP, OCP, DMTF, Redfish, or other open-source platform-management and networking projects!
  • Experience building a switch network operating system, BMC-managed switch subsystem, or converged server/switch management architecture for hyperscale or enterprise platforms.

Expected Impact

  • Deliver a reusable OpenBMC switch-management architecture that supports multiple switch backend implementations without tying the platform model to one ASIC vendor.
  • Enable a unified management experience for host platform health, BMC management, and ToR switch network health through coherent inventory, configuration, telemetry, and lifecycle workflows.
  • Improve platform reliability by providing safe defaults, validated configuration persistence, rollback, diagnostics, and strong separation between management and data-plane traffic.