Data Center & AI Infrastructure Engineer
Open a video or resource link, watch it, then check its box to track your progress and unlock what comes next.
- Engineer power, cooling and rack-scale accelerated compute
- Build high-bandwidth AI fabrics + production GPU clusters
- Operate resilient, efficient AI factories at scale
Benchmark: OpenAI Compute / Data Center Engineering + Google Data Center Technology & Systems + Meta infrastructure signals (current Sep 2026) Market refresh: high-density liquid-cooled racks; power/thermal architecture; rack-scale GPUs; AI fabrics; cluster provisioning; fleet health; commissioning, digital twins, capacity and efficiency.
Who’s Hiring Right Now
US openings posted in the last 7 days at OpenAI — this role’s own benchmark employers — refreshed daily, as of 2026-11-01. Pulled directly from each company’s public job board; not exhaustive, and postings close fast — worth confirming the details on the employer’s own site before applying.
Engineer power, cooling and rack-scale accelerated compute
AI data-center power + liquid cooling
Coverage: C168 Data Center Engineering
Rack-scale GPU systems + server topology
Coverage: C152 / C126
Build high-bandwidth AI fabrics + production GPU clusters
AI networking, NVLink/NCCL + Ethernet fabrics
Coverage: C130 / C126
Kubernetes GPU lifecycle + cluster operations
Coverage: C080 / C068 / C126
Operate resilient, efficient AI factories at scale
Fleet health, observability + failure recovery
Coverage: C083 / C126
Digital twins, commissioning + capacity efficiency
Coverage: C168 / C083
Role-Readiness Proof / Capstone
Design and validate a production-ready AI data-center pod/campus slice: define GPU/rack and workload assumptions; power one-line, redundancy, rack density and liquid-cooling/CDU architecture; specify Ethernet/InfiniBand/NVLink fabrics, provisioning/scheduling and storage paths; build capacity, telemetry/SLO and failure-domain models; create FAT/SAT + integrated commissioning tests and a digital-twin scenario model; quantify PUE/WUE, usable GPU capacity, bottlenecks, cost and recovery tradeoffs.
Proof Rubric
- Show engineering evidence, not a diagram: defend power, cooling, rack and fabric assumptions
- demonstrate capacity and failure-domain math
- validate collective/network performance
- automate cluster bring-up and GPU lifecycle
- prove telemetry, SLOs, health checks and incident diagnosis
- execute commissioning/failure tests
- quantify PUE/WUE, GPU utilization, availability and cost
- explain residual risks, maintainability and scale-out tradeoffs