AI Data Center Management
Operate AI infrastructure safely and efficiently through observability, capacity control, reliability engineering, security, automation, and continuous improvement.
The AI Data Center Operating System
Learning objective
By the end of this lesson, the learner will be able to explain the ai data center operating system, apply it to a realistic situation, and defend the chosen action with evidence.
Core lesson
Align facilities, network, platform, security, finance, and customer operations around service objectives, clear ownership, change control, and a shared operating picture.
Instructor teaching guidance
Begin by establishing what the learner already knows about the ai data center operating system. Teach the principle in plain language, define unfamiliar terms, then connect it to the course goal. Ask the learner to explain the idea back in their own words before moving to application.
Applied exercise
Create a RACI for an AI cluster spanning facilities, platform, network, security, and vendor teams.
Facilitation sequence
Guide the learner through the application in five steps: describe the present situation, identify the lesson principle, choose an action, name the evidence to review, and define how success will be measured. Use follow-up questions when an answer is vague or unsupported.
Mastery standard
The operating model must manage the whole service, not isolated technical silos.
Knowledge check
Ask the learner to explain why “The operating model must manage the whole service, not isolated technical silos.” is the correct standard, then identify one limitation, one risk, and one piece of evidence that could change the answer. Do not mark mastery from confidence alone; require clear reasoning and practical application.
Day 0, Day 1, and Day 2 Operations
Learning objective
By the end of this lesson, the learner will be able to explain day 0, day 1, and day 2 operations, apply it to a realistic situation, and defend the chosen action with evidence.
Core lesson
Day 0 establishes standards and readiness; Day 1 deploys and validates; Day 2 sustains, optimizes, patches, expands, and retires infrastructure.
Instructor teaching guidance
Begin by establishing what the learner already knows about day 0, day 1, and day 2 operations. Teach the principle in plain language, define unfamiliar terms, then connect it to the course goal. Ask the learner to explain the idea back in their own words before moving to application.
Applied exercise
Build a readiness checklist for receiving a new liquid-cooled rack-scale system.
Facilitation sequence
Guide the learner through the application in five steps: describe the present situation, identify the lesson principle, choose an action, name the evidence to review, and define how success will be measured. Use follow-up questions when an answer is vague or unsupported.
Mastery standard
Operational excellence begins before equipment arrives.
Knowledge check
Ask the learner to explain why “Operational excellence begins before equipment arrives.” is the correct standard, then identify one limitation, one risk, and one piece of evidence that could change the answer. Do not mark mastery from confidence alone; require clear reasoning and practical application.
Observability, DCIM, Telemetry, and Digital Twins
Learning objective
By the end of this lesson, the learner will be able to explain observability, dcim, telemetry, and digital twins, apply it to a realistic situation, and defend the chosen action with evidence.
Core lesson
Correlate electrical, thermal, water, network, storage, GPU, scheduler, security, and application telemetry. Use baselines, event context, topology, and digital models to convert signals into decisions.
Instructor teaching guidance
Begin by establishing what the learner already knows about observability, dcim, telemetry, and digital twins. Teach the principle in plain language, define unfamiliar terms, then connect it to the course goal. Ask the learner to explain the idea back in their own words before moving to application.
Applied exercise
Design one dashboard that connects facility conditions to job performance and business impact.
Facilitation sequence
Guide the learner through the application in five steps: describe the present situation, identify the lesson principle, choose an action, name the evidence to review, and define how success will be measured. Use follow-up questions when an answer is vague or unsupported.
Mastery standard
A wall of metrics is not observability; teams need context, causality, and action thresholds.
Knowledge check
Ask the learner to explain why “A wall of metrics is not observability; teams need context, causality, and action thresholds.” is the correct standard, then identify one limitation, one risk, and one piece of evidence that could change the answer. Do not mark mastery from confidence alone; require clear reasoning and practical application.
Capacity and Workload Orchestration
Learning objective
By the end of this lesson, the learner will be able to explain capacity and workload orchestration, apply it to a realistic situation, and defend the chosen action with evidence.
Core lesson
Manage power, cooling, floor space, accelerators, network, storage, quotas, reservations, batch windows, priority, and fragmentation across training and inference demand.
Instructor teaching guidance
Begin by establishing what the learner already knows about capacity and workload orchestration. Teach the principle in plain language, define unfamiliar terms, then connect it to the course goal. Ask the learner to explain the idea back in their own words before moving to application.
Applied exercise
Resolve a scenario where three teams request the same constrained GPU pool.
Facilitation sequence
Guide the learner through the application in five steps: describe the present situation, identify the lesson principle, choose an action, name the evidence to review, and define how success will be measured. Use follow-up questions when an answer is vague or unsupported.
Mastery standard
Capacity management balances service priority, utilization, fairness, resilience, and future headroom.
Knowledge check
Ask the learner to explain why “Capacity management balances service priority, utilization, fairness, resilience, and future headroom.” is the correct standard, then identify one limitation, one risk, and one piece of evidence that could change the answer. Do not mark mastery from confidence alone; require clear reasoning and practical application.
Thermal and Liquid-Cooling Operations
Learning objective
By the end of this lesson, the learner will be able to explain thermal and liquid-cooling operations, apply it to a realistic situation, and defend the chosen action with evidence.
Core lesson
Monitor supply and return temperatures, flow, pressure, approach temperature, water quality, leak detection, CDU health, dew point, and rack-level heat rejection. Maintain safe transition states.
Instructor teaching guidance
Begin by establishing what the learner already knows about thermal and liquid-cooling operations. Teach the principle in plain language, define unfamiliar terms, then connect it to the course goal. Ask the learner to explain the idea back in their own words before moving to application.
Applied exercise
Respond to rising coolant return temperature without causing a workload or condensation incident.
Facilitation sequence
Guide the learner through the application in five steps: describe the present situation, identify the lesson principle, choose an action, name the evidence to review, and define how success will be measured. Use follow-up questions when an answer is vague or unsupported.
Mastery standard
Thermal operations require coordinated facility controls, hardware limits, workload response, and practiced escalation.
Knowledge check
Ask the learner to explain why “Thermal operations require coordinated facility controls, hardware limits, workload response, and practiced escalation.” is the correct standard, then identify one limitation, one risk, and one piece of evidence that could change the answer. Do not mark mastery from confidence alone; require clear reasoning and practical application.
Power Quality and Energy Operations
Learning objective
By the end of this lesson, the learner will be able to explain power quality and energy operations, apply it to a realistic situation, and defend the chosen action with evidence.
Core lesson
Track utility conditions, demand, UPS state, batteries, generators, busways, rack power, harmonics, transients, and workload-driven ramps. Coordinate demand response only within resilience limits.
Instructor teaching guidance
Begin by establishing what the learner already knows about power quality and energy operations. Teach the principle in plain language, define unfamiliar terms, then connect it to the course goal. Ask the learner to explain the idea back in their own words before moving to application.
Applied exercise
Build an operating response to a forecast grid constraint during a long training run.
Facilitation sequence
Guide the learner through the application in five steps: describe the present situation, identify the lesson principle, choose an action, name the evidence to review, and define how success will be measured. Use follow-up questions when an answer is vague or unsupported.
Mastery standard
Energy optimization never outranks life safety, equipment protection, and committed service reliability.
Knowledge check
Ask the learner to explain why “Energy optimization never outranks life safety, equipment protection, and committed service reliability.” is the correct standard, then identify one limitation, one risk, and one piece of evidence that could change the answer. Do not mark mastery from confidence alone; require clear reasoning and practical application.
Reliability Engineering and Maintenance
Learning objective
By the end of this lesson, the learner will be able to explain reliability engineering and maintenance, apply it to a realistic situation, and defend the chosen action with evidence.
Core lesson
Use criticality, failure modes, condition-based maintenance, spares, firmware governance, maintenance windows, bypass controls, and post-maintenance validation.
Instructor teaching guidance
Begin by establishing what the learner already knows about reliability engineering and maintenance. Teach the principle in plain language, define unfamiliar terms, then connect it to the course goal. Ask the learner to explain the idea back in their own words before moving to application.
Applied exercise
Create a maintenance plan for a component with no safe concurrent bypass.
Facilitation sequence
Guide the learner through the application in five steps: describe the present situation, identify the lesson principle, choose an action, name the evidence to review, and define how success will be measured. Use follow-up questions when an answer is vague or unsupported.
Mastery standard
Reliable maintenance requires explicit risk controls before, during, and after the work.
Knowledge check
Ask the learner to explain why “Reliable maintenance requires explicit risk controls before, during, and after the work.” is the correct standard, then identify one limitation, one risk, and one piece of evidence that could change the answer. Do not mark mastery from confidence alone; require clear reasoning and practical application.
Incident Command and Root-Cause Analysis
Learning objective
By the end of this lesson, the learner will be able to explain incident command and root-cause analysis, apply it to a realistic situation, and defend the chosen action with evidence.
Core lesson
Define severity, command roles, technical workstreams, communications, decision logs, recovery criteria, evidence preservation, and blameless learning. Separate proximate triggers from systemic causes.
Instructor teaching guidance
Begin by establishing what the learner already knows about incident command and root-cause analysis. Teach the principle in plain language, define unfamiliar terms, then connect it to the course goal. Ask the learner to explain the idea back in their own words before moving to application.
Applied exercise
Run a tabletop exercise for simultaneous cooling alarms and GPU job failures.
Facilitation sequence
Guide the learner through the application in five steps: describe the present situation, identify the lesson principle, choose an action, name the evidence to review, and define how success will be measured. Use follow-up questions when an answer is vague or unsupported.
Mastery standard
Fast recovery and deep learning require a practiced incident structure.
Knowledge check
Ask the learner to explain why “Fast recovery and deep learning require a practiced incident structure.” is the correct standard, then identify one limitation, one risk, and one piece of evidence that could change the answer. Do not mark mastery from confidence alone; require clear reasoning and practical application.
AI-Assisted Operations and Automation
Learning objective
By the end of this lesson, the learner will be able to explain ai-assisted operations and automation, apply it to a realistic situation, and defend the chosen action with evidence.
Core lesson
Apply anomaly detection, forecasting, predictive maintenance, optimization, copilots, and closed-loop automation with evidence, confidence thresholds, human override, audit logs, and safe rollback.
Instructor teaching guidance
Begin by establishing what the learner already knows about ai-assisted operations and automation. Teach the principle in plain language, define unfamiliar terms, then connect it to the course goal. Ask the learner to explain the idea back in their own words before moving to application.
Applied exercise
Classify five automations as advisory, approval-required, or bounded autonomous.
Facilitation sequence
Guide the learner through the application in five steps: describe the present situation, identify the lesson principle, choose an action, name the evidence to review, and define how success will be measured. Use follow-up questions when an answer is vague or unsupported.
Mastery standard
Automate within a verified safety envelope and keep accountable humans able to intervene.
Knowledge check
Ask the learner to explain why “Automate within a verified safety envelope and keep accountable humans able to intervene.” is the correct standard, then identify one limitation, one risk, and one piece of evidence that could change the answer. Do not mark mastery from confidence alone; require clear reasoning and practical application.
Zero Trust and Platform Security
Learning objective
By the end of this lesson, the learner will be able to explain zero trust and platform security, apply it to a realistic situation, and defend the chosen action with evidence.
Core lesson
Continuously verify users, devices, workloads, services, and administrative actions. Protect secrets, images, firmware, APIs, management networks, model artifacts, and training data.
Instructor teaching guidance
Begin by establishing what the learner already knows about zero trust and platform security. Teach the principle in plain language, define unfamiliar terms, then connect it to the course goal. Ask the learner to explain the idea back in their own words before moving to application.
Applied exercise
Design a privileged-access workflow for an emergency vendor intervention.
Facilitation sequence
Guide the learner through the application in five steps: describe the present situation, identify the lesson principle, choose an action, name the evidence to review, and define how success will be measured. Use follow-up questions when an answer is vague or unsupported.
Mastery standard
Network location alone must never confer trust.
Knowledge check
Ask the learner to explain why “Network location alone must never confer trust.” is the correct standard, then identify one limitation, one risk, and one piece of evidence that could change the answer. Do not mark mastery from confidence alone; require clear reasoning and practical application.
SLOs, FinOps, Sustainability, and Reporting
Learning objective
By the end of this lesson, the learner will be able to explain slos, finops, sustainability, and reporting, apply it to a realistic situation, and defend the chosen action with evidence.
Core lesson
Connect availability, job completion, latency, utilization, energy, water, carbon, cost, forecast accuracy, and customer experience into balanced operating objectives.
Instructor teaching guidance
Begin by establishing what the learner already knows about slos, finops, sustainability, and reporting. Teach the principle in plain language, define unfamiliar terms, then connect it to the course goal. Ask the learner to explain the idea back in their own words before moving to application.
Applied exercise
Create an executive scorecard with leading indicators and action thresholds.
Facilitation sequence
Guide the learner through the application in five steps: describe the present situation, identify the lesson principle, choose an action, name the evidence to review, and define how success will be measured. Use follow-up questions when an answer is vague or unsupported.
Mastery standard
Optimize useful AI output and service value, not a single isolated metric.
Knowledge check
Ask the learner to explain why “Optimize useful AI output and service value, not a single isolated metric.” is the correct standard, then identify one limitation, one risk, and one piece of evidence that could change the answer. Do not mark mastery from confidence alone; require clear reasoning and practical application.
Capstone: Operate Through a Critical Week
Learning objective
By the end of this lesson, the learner will be able to explain capstone: operate through a critical week, apply it to a realistic situation, and defend the chosen action with evidence.
Core lesson
Integrate staffing, maintenance, capacity contention, grid alerts, security events, customer priorities, and executive communication into one controlled operational plan.
Instructor teaching guidance
Begin by establishing what the learner already knows about capstone: operate through a critical week. Teach the principle in plain language, define unfamiliar terms, then connect it to the course goal. Ask the learner to explain the idea back in their own words before moving to application.
Applied exercise
Lead a simulated week containing a coolant leak alarm, urgent inference launch, firmware advisory, and utility curtailment request.
Facilitation sequence
Guide the learner through the application in five steps: describe the present situation, identify the lesson principle, choose an action, name the evidence to review, and define how success will be measured. Use follow-up questions when an answer is vague or unsupported.
Mastery standard
World-class operations are calm, evidence-led, rehearsed, transparent, and continuously improving.
Knowledge check
Ask the learner to explain why “World-class operations are calm, evidence-led, rehearsed, transparent, and continuously improving.” is the correct standard, then identify one limitation, one risk, and one piece of evidence that could change the answer. Do not mark mastery from confidence alone; require clear reasoning and practical application.