Power Architect - Data Center Infrastructure

STEPS, INC
Plano, TX, United States
7 days ago
Apply on www.dice.com
Prepare application

Role details

Contract type
Permanent contract
Employment type
Full-time (> 32 hours)
Experience level
Expert
Experience required
10 years minimum
Working hours
Regular working hours
Job source

Tech stack

Artificial Intelligence Systems Engineering Computer Engineering Data Centers Ethernet InfiniBand PCI Express Nvme

Job description

Our client is seeking a Power Architect - Data Center Infrastructure to lead the development of next-generation Grid-to-Chip power delivery architectures supporting hyperscale and enterprise data center platforms.

This role will drive system-level power strategy across facility, rack, and board levels, with a focus on efficient, scalable, and reliable power solutions for rapidly increasing AI/HPC power requirements.

The ideal candidate brings deep experience in data center power infrastructure, power conversion, distribution, and system-level architecture., * Define and drive end-to-end data center power architecture spanning the utility/grid interface, UPS systems, power distribution, rack-level power shelves, and board-level voltage regulation.

  • Lead next-generation Grid-to-Chip power delivery strategies for hyperscale and AI/HPC data center platforms.
  • Architect high-efficiency power conversion and distribution systems for modern compute platforms.
  • Translate AI, cloud, and HPC workload trends into facility-, row-, and rack-level power requirements.
  • Develop reference power architectures and system design guidelines.
  • Evaluate and optimize AC-DC and DC-DC power conversion topologies, including centralized, distributed, and intermediate bus architectures.
  • Lead trade-off analysis across efficiency, scalability, reliability, and cost.
  • Develop architectural strategies supporting 100kW+ racks, AI clusters, and advanced accelerator platforms.
  • Develop proof-of-concept systems and reference designs with internal and external engineering teams.
  • Collaborate with system engineering, product management, silicon partners, hyperscale customers, and ecosystem partners.
  • Provide technical leadership and influence product roadmaps for evolving hyperscale infrastructure requirements.

Requirements

  • Deep knowledge of data center power architecture and infrastructure.
  • Experience designing Grid-to-Chip power delivery systems.
  • Expertise in AC-DC and DC-DC conversion architecture.
  • Experience with UPS systems, PDUs, rack power shelves, and busbar architecture.
  • Strong understanding of high-power rack architectures for hyperscale and AI workloads.
  • Familiarity with server and accelerator platform power delivery requirements.
  • Understanding of CPU/GPU compute platforms and their impact on power delivery.
  • Knowledge of technologies including PCIe, CXL, Ethernet, InfiniBand, and NVMe/NVMe-oF.
  • Ability to evaluate system-level power efficiency and reliability trade-offs., * BS, MS, or PhD in Electrical Engineering, Computer Engineering, Mechanical Engineering, or a related discipline.
  • 10+ years of experience in power architecture or power systems design.
  • Extensive experience designing power infrastructure for data center platforms.
  • Demonstrated expertise in Grid-to-Chip power architecture and high-power rack systems.
  • Strong background in power electronics and system-level power delivery design.
  • Experience influencing architecture decisions across cross-functional engineering teams.
  • Experience working with hyperscale data center operators, server OEMs, or infrastructure vendors.

Apply for this position

This job is hosted externally. Click below to view the full posting and apply.

Apply on www.dice.com
Prepare application

Good distractions

Talks and stories from around this role — technically off-topic, practically not.

1:29 min

Tech infrastructure capacity and AI product innovations

1:57 min

Routing cross-rack traffic seamlessly with NCCL

Kevin Klues Kevin Klues · World Congress 2025

4:52 min

Connecting namespaces with local virtual ethernet pairs

Oliver Seitz Oliver Seitz · World Congress 2025

1:51 min

Bypassing the CPU stack with remote direct memory access

Lerna Ekmekcioglu Lerna Ekmekcioglu · Europe 2026 Virtual

3:23 min

The AI workload technology stack and its components

Lerna Ekmekcioglu Lerna Ekmekcioglu · Europe 2026 Virtual

2:12 min

Implementing automotive ethernet and connected remote vehicle telemetry applications

David Romić · World Congress 2023

Videos

See all

Related articles

See all