AI Data Center Power Architect

BIZLINK INTERCONNECT TECHNOLOGY ( INDIA ) PRIVATE LIMITED; Fremont (CA); On-site; USD 150,000 - 260,000; Full time

Job summary

BIZLINK INTERCONNECT TECHNOLOGY ( INDIA ) PRIVATE LIMITED seeks a senior power architect to design and optimize high-density power delivery for AI servers and racks. You will develop models, validate performance, and ensure compliance with IEC/IEEE and thermal standards across global data-center deployments.

Lead collaboration with OEMs and internal teams to drive innovations in liquid-cooled busbars and advanced materials, while representing the company in industry consortia and standards

Qualifications

• Bachelor’s degree in Electrical/Mechanical Engineering or related field.
• 10–15 years of experience in power architecture for high-density AI infrastructure.
• Expertise in high-current designs up to 3,000A and liquid-cooled busbars.
• Experience with IEC/IEEE standards and CAD tools.
• Strong cross-functional collaboration with OEMs and data-center operators.

Responsibilities

• Design and optimize power distribution systems for AI servers.
• Develop power estimation models and tools for rack-level systems.
• Ensure compliance with IEC/IEEE standards and seismic/thermal requirements.
• Collaborate with sales, PLM, and R&D to align roadmaps.
• Represent organization in industry consortia (OCP, IEEE).

Skills

Power architecture

High-current design

Thermal management

CAD tools

Standards compliance

OEM collaboration

Reference architectures

Communication skills

AI server power delivery

Education

Bachelor’s degree in Electrical/Mechanical Engineering or related field

Master’s/PhD in Power Mechanics/ Thermal Engineering/ Applied Physics

Tools

Power modeling tools

Simulation tools

Job description

• Design & Optimization: Architect power distribution systems (connectors, cables, busbars, whips) for AI servers, ensuring scalability, reliability, and thermal efficiency.
• Focus on high-current applications (e.g., 130–3,000A) and alignment precision (±1.00mm) to prevent damage in dense GPU/TPU environments.
• Power Estimation and Modeling:
• Build and maintain accurate power estimation models and tools for AI server and rack-level systems.
• Conduct performance versus power analyses, support early architectural exploration, and collaborate with hardware/software teams to improve Perf/Watt metrics and total cost of ownership for next-generation AI platforms
• Compliance & Standards:
• Ensure designs meet IEC 61439, IEEE, and safety regulations, including seismic and thermal management requirements (e.g., -40°C to +125°C operating range).
• Cross-Functional Collaboration: Partner with sales, PLM, and R&D to align product roadmaps with AI infrastructure demands (e.g., NVIDIA GPU power needs) and customer requirements.
• Innovation & Trends: Integrate emerging technologies like liquid-cooled busbars, low-GWP cooling, and advanced materials (high-conductivity copper alloys) to enhance power density and efficiency.
• Industry Engagement: Represent the organization in consortia like OCP, OIF, and IEEE to drive standardization and partnerships, leveraging reference architectures for rapid deployment.
• Minimum: Bachelor’s degree in Electrical Engineering, Mechanical Engineering, Power Systems, or related field.
• Preferred: Master’s/PhD in Power Mechanics, Thermal Engineering, or Applied Physics
• Over 10–15 years of experience in power architecture for high-density systems (100kW-1MW/rack), including AI/ML infrastructure.
• Expertise in power connector/busbar design (e.g., liquid-cooled, solder-tail, screw-mount), high-current systems (up to 3,000A), and thermal management such as liquid-cooling for AI workloads.
• Proficiency in IEC/IEEE standards, CAD tools, and system-level validation (e.g., 25kA short-circuit rating).
• Experience collaborating with OEMs, chip vendors (NVIDIA, AMD, Intel), and data center operators to optimize AI server and rack power delivery.
• Strong communication skills to bridge technical teams and business stakeholders.
• Knowledge of dynamic workload management (e.g., BBU/powershelf).
• Familiarity with AI reference architectures (e.g., NVIDIA designs supporting 132 kW/rack).