AI Data Center Manager

About the role

SoftConstruct is seeking an experienced, forward‑thinking AI Data Center Manager to lead a 5 MW high‑density compute footprint dedicated to enterprise AI workloads such as large‑language‑model fine‑tuning, high‑throughput inference and multi‑node HPC clustering. The role combines physical deployment, thermal management and day‑to‑day operational efficiency of liquid‑cooled GPU clusters.

Key responsibilities

• Oversee operation and maintenance of direct‑to‑chip liquid‑cooling loops, Coolant Distribution Units (CDU) and rear‑door heat exchangers (RDHx) in a dense environment.
• Manage rack‑level power distribution for ultra‑dense power envelopes up to 60 kW per rack, ensuring stable delivery across the 5 MW footprint.
• Coordinate with colocation partners and internal AI Platform Engineering teams to guarantee maximum hardware performance and zero thermal throttling.
• Lead the transition from legacy air‑cooled to hybrid or full liquid‑cooled architectures.
• Negotiate and monitor Service Level Agreements (SLAs), Master Service Agreements (MSAs) and joint operating procedures with commercial colocation facilities.

Required profile

• 7+ years of experience in data‑center operations, high‑performance computing or critical‑facilities engineering, with at least 3 years in a leadership role.
• Bachelor’s degree in Mechanical Engineering, Electrical Engineering, Computer Science or equivalent practical experience.
• Hands‑on experience deploying and maintaining high‑power‑density GPU hardware such as NVIDIA HGX/DGX and liquid‑cooled chassis.
• Deep technical knowledge of MEP design and direct‑to‑chip liquid cooling systems.
• Proven track record negotiating colocation contracts and managing SLAs.

Required skills

• Direct‑to‑chip liquid cooling loops
• Coolant Distribution Units (CDU) and rear‑door heat exchangers (RDHx)
• High‑density GPU hardware (NVIDIA HGX, DGX)
• Kubernetes container orchestration
• Slurm workload scheduler