AI Data Center Engineering Professional Training Program

Why AI/GPU data centers are a distinct engineering discipline — power, cooling, and network fabric as one interacting system.

NFPA 76, ASHRAE TC9.9, Uptime Institute Tiers, and NEC Article 645.

Utility service through UPS, PDUs, and busway distribution to the rack.

Rear-door heat exchangers, direct-to-chip, immersion cooling, and CDU architecture.

InfiniBand vs. RoCE, leaf-spine topology, and physical layout implications.

From GPU server specs to a completed PDU circuit schedule.

Worked examples — UPS sizing, voltage drop, and generator step-load.

Worked examples — cooling tons, liquid/air split, CDU flow rate.

Utility interconnection studies and on-site generation as a capacity bridge.

N+1 vs. 2N architecture and the cost/availability tradeoff.

One-line diagrams, cooling P&IDs, rack elevations, and white-space plans.

Field practices, leak-detection installation, and load-bank testing.

DCIM point design, real-time PUE/WUE, and capacity planning data.

9 real diagnostic scenarios with symptoms, diagnosis, and resolution.

PUE, WUE, and CUE calculation, and building a defensible efficiency report.

Retrofitting a 5 MW enterprise colo to support a 60 kW/rack liquid-cooled AI tenant.

A 150 MW greenfield AI training campus with on-site substation and generation.

A 500 kW edge inference facility co-located at a telecom central office.

Converting an air-cooled 2 MW hall to direct-to-chip liquid cooling.

A modular 1 MW facility using prefabricated power/cooling skids.

An 800 kW research cluster built into existing campus infrastructure.

A 300 kW private cluster for proprietary training inside corporate IT space.

A purpose-built multi-tenant facility with shared central utility plant.

A downloadable kit of 12 real project templates.

A final design assignment, a comprehensive exam, and your certificate of completion.