Senior MEP Design Engineer - HPC​/AI Data Centers

About Submer

Submer designs and delivers end-to-end AI datacenter infrastructure built around industry-leading liquid cooling.

We help organizations scale AI beyond the limits of traditional datacenters by enabling higher density, greater efficiency and lower environmental impact – accelerating time-to-AI from first deployment to full production.

We are seeking an MEP Design Engineer - HPC/AI Data Centers to lead the full Mechanical, Electrical, and Plumbing (MEP) design and delivery of high-performance computing (HPC) and AI data centers. This role owns the integrated MEP architecture for large-scale 100-300 MW facilities and modular data center (MDC) deployments that scale into 50-300 MW HPC/AI sites. You will ensure seamless coordination across mechanical, electrical, and plumbing systems to support next-generation GPU clusters, with a strong emphasis on NVIDIA Grace Blackwell (GB300) and Vera Rubin (VR200) reference architectures.

This is a hands‑on technical leadership position ideal for an engineer who can own end-to-end MEP design for mission‑critical, high‑density environments.

• Lead the holistic MEP design, specification, and optimization for hyperscale and modular HPC/AI data centers, ensuring integrated performance, reliability, efficiency, and scalability.
• Design electrical systems including MV/LV distribution, switchboards, overhead busways, circuit breakers, transformers (ML/VL), PDUs, grounding, protection, BESS, UPS, and generators.
• Oversee mechanical systems focused on high-density liquid cooling (direct-to-chip, immersion, CDUs, in-row/in-rack solutions), CRAC/CRAH, heat rejection, airflow management, and redundancy for 100+ kW racks.
• Manage plumbing and fluid systems, including coolant distribution, primary/secondary cooling loops, leak detection, piping, pumps, filtration, and integration with facility water systems or closed-loop designs.
• Architect modular data center solutions (1‑2 MW containers/pods) that efficiently aggregate into larger 50-300 MW HPC/AI sites, emphasizing standardization, rapid deployment, and future-proofing.
• Deliver MEP designs tailored to NVIDIA GB300 (Grace Blackwell) and VR200 (Vera Rubin) Superpod/reference architectures, addressing rack power (~120 kW+), liquid cooling manifolds, power smoothing, high-current delivery, and thermal management requirements.
• Perform system-level studies (power, thermal, hydraulic, CFD) and ensure compliance with NEC, NFPA, ASHRAE, IEEE, Uptime Institute, and local codes.
• Collaborate with architecture, structural, IT, and construction teams throughout design, procurement, construction, commissioning, and handover phases.
• Optimize for energy efficiency, sustainability, capacity planning, fault tolerance, and total cost of ownership in AI factory environments.
• Develop standards, create and review drawings/specifications, troubleshoot issues, and mentor junior MEP engineers.
• Create internal and customer-facing MEP specifications and explain/discuss with technical and non-technical customers.

• Bachelor’s degree (Master’s preferred) in Mechanical Engineering, Electrical Engineering, or a related MEP-focused discipline. PE license strongly preferred.
• 5+ years of progressive experience in MEP engineering for mission-critical facilities, with at least 3‑5 years in large-scale data centers or equivalent high-density environments.
• Demonstrated ability to design complete HPC/AI data centers across all three MEP disciplines (Mechanical, Electrical, Plumbing/Cooling).
• Proven track record with HPC/AI data centers at 100-300 MW scale and modular deployments (1‑2 MW units scaling to 50-300 MW sites).
• Strong knowledge of NVIDIA Grace Blackwell (GB300) and Vera Rubin (VR200) architectures, including integrated power, cooling, and fluid system requirements.
• Expertise in high-density liquid cooling, Grid-to-Rack power distribution, all MEP equipment in single-line power specification, 13.8kV/34.5kV to IT white spaces with 480V AC and/or 800V DC power distribution, BESS integration, MV/LV power distribution, busways, and critical infrastructure redundancy (N+1, 2N, etc.).
• Proficiency with relevant…