Posted 6 weeks ago

Software Developer 3

Taipei, Taiwan 🇹🇼
  • Onsite
  • Python
  • Terraform

Job description

Scaled Manufacturing (SMF) builds and operates container-based services for validating infrastructure across factory environments. SMF enables repeatable production of liquid-cooled GPU platforms - including GB200, GB300, VR, MI355, and MI455 - by embedding quality, yield, throughput, and readiness controls into manufacturing. SMF delivers factory infrastructure; a Control Plane for test lifecycle management, operator workflows, quality gates, fleet health, and reporting; and a Data Plane that automates tests validation. Manufacturing Repair and Triage identifies root causes, guides recovery and uses data to reduce repeat failures. Working with Supply Chain Operations, Oracle Hardware Development, manufacturing partners, data-center teams, NVIDIA, and AMD, SMF detects hardware, firmware, software, and configuration issues before racks ship. This improves yield, reduces rework and downstream failures, and accelerates reliable AI infrastructure deployment at hyperscale.

GPU Platform Validation and Provisioning

  • Develop automated workflows for GPU server provisioning, configuration, firmware installation, validation, upgrades, and recovery.
  • Build and maintain GPU burn-in, qualification, and factory acceptance tests. Define pass/fail criteria and document results for each approved platform configuration.
  • Validate GPU health, device enumeration, memory, PCIe and GPU interconnects, networking, power, cooling, and performance under sustained workloads.
  • Use platform-specific diagnostics, including NVIDIA DCGM and CUDA-based tools and AMD ROCm Validation Suite (RVS) and AMD SMI.

Failure Analysis and Manufacturing Readiness

  • Investigate failures across GPUs, server hardware, BIOS/UEFI, BMC, firmware, Linux kernels, drivers, operating systems, storage, and networking.
  • Capture logs, test results, firmware and software versions, and component identifiers to support traceability and root cause analysis.
  • Work with ODM engineers and internal and external partners to verify corrective actions and prevent repeat failures.
  • Use manufacturing yield, test throughput, failure trends, and quality metrics to improve test coverage and deployment readiness.

Infrastructure and Networking:

  • Network engineering, data-center/site networking, VLANs, VxLAN firewalls, L3 networking, gNMI, VPNs, DHCP, switch configuration (Arista), network automation, troubleshooting, Linux, Python.
  • Host provisioning/configuration/Terraform, containers, monitoring/observability, Prometheus, Grafana, Linux, Python
  • Experience with Identity Provider tools such as Keycloak and FreeIPA,

Automation and Collaboration

  • Develop reliable validation tooling that can run across large factory fleets and report system health and test outcomes.
  • Partner with provisioning, platform, and data center teams to ensure factory tests reflect deployment requirements.
  • Document test procedures, known issues, recovery steps, and production release recommendations.

Minimum Job Qualifications

  • Experience bringing up, validating, or troubleshooting GPU, accelerator, or server hardware in a lab, manufacturing, or data center environment.
  • Strong Linux skills, including operating system installation and imaging, kernel versions and modules, device drivers, boot processes, system logs, and hardware troubleshooting.
  • Ability to diagnose failures across the hardware, firmware, operating system, kernel, and driver layers.
  • Knowledge of CPU and GPU architecture, PCIe, device enumeration, BIOS/UEFI, BMC interfaces, and power and thermal behavior.
  • Networking knowledge, including Linux network configuration, Smart NICs , IP addressing, DHCP/DNS, and troubleshooting network connectivity. Familiarity with high-performance networking concepts such as RDMA or RoCE.
  • Experience scripting and automating validation or provisioning workflows using Python and shell.
  • Ability to develop dependable test workflows for large-scale systems, analyze results across many servers, and identify recurring failures or performance outliers.
  • Experience documenting test plans, failure investigations, and results while working across engineering and manufacturing teams.
  • Strong experience in data pipelines, data modeling, analytics and measurement systems, and applied AI pipelines.

Preferred Job Qualifications

  • Hands-on experience with both NVIDIA CUDA/DCGM and AMD ROCm/RVS environments.
  • Experience validating multi-GPU systems, GPU interconnects, PCIe bandwidth, and single-node or multi-node workloads.
  • Familiarity with PXE/iPXE provisioning, Redfish, PLDM, virtual media, firmware lifecycle management, and factory test infrastructure.
  • Experience with component traceability, manufacturing yield analysis, and corrective-action processes.
  • Ability to work on site at ODM facilities during platform bring-up and production ramps.

Other jobs in Taiwan

ASIC Engineer

Cisco Zhubei, Taiwan🇹🇼

Hybrid Full Time 22 hours ago