Home Dashboard
Tickets and Groupings
Floor Plan and Optimal Route
3D Model of Server
Inventory

DC Daddy - NMC DataCenter Command Center

Inspiration

Data centers are the backbone of modern computing, yet technicians managing $2.8 billion facilities powering 400MW of compute still rely on fragmented tools and manual processes. We were inspired by NMC's challenge of reducing human error and operational inefficiency in environments where a single mistake can cost millions. We saw an opportunity to build an integrated command center that transforms how data center operations work, making technicians' jobs smarter, not harder, while prioritizing their health and safety.

What it does

DC Daddy is a comprehensive data center operations platform that solves five critical pain points through an integrated ecosystem:

1. Intelligent Work Order Management

Multi-Channel Integration: Automatically creates tickets from emails, Slack messages, and manual input using AI-powered parsing
Dual-Mode Views:
- Admin Mode: Focus on customer impact, severity, and one-click automation (provisioning, permissions)
- Technician Mode: Optimized task routing with distance, floor elevation, and priority
Smart Bundling: Groups parallelizable tickets to save time and reduce travel
Health-Based Task Assignment: Automatically flags physically demanding tasks (L2/L3 floors) for technicians with health conditions (knee pain, arthritis, etc.)

2. AI-Powered Communication Hub

Unstructured Text Parsing: Extracts structured ticket information from emails and Slack
Automatic Categorization: Identifies severity, type, customer impact, and routing details
Zero Manual Entry: Tickets are auto-generated and routed to correct teams
Real-time Updates: Communication feed shows auto-generated updates across channels

3. Optimized Technician Routing

Smart Pathfinding: Groups tasks by proximity, floor, and priority
Hot/Cold Zone Rotation: Alternates between server zones to prevent technician fatigue
Elevation-Aware: Considers floor changes in route optimization
Visual Floor Plans: Interactive maps showing task locations and optimal routes
Health Considerations: Routes avoid excessive physical strain for technicians with health conditions

4. Predictive Inventory Management

Real-time Tracking: Monitor stock levels across all data center locations
Predictive Depletion: Calculates weeks remaining based on usage trends with mathematical forecasting
Color-Coded Alerts: Visual status indicators (Healthy, Warning, Critical, Depleted)
Advanced Filtering: Search by name, category, site; filter by status (in-use, spare, idle, drained)
Usage Trend Analysis: Tracks rising, steady, or cooling consumption patterns
Threshold Warnings: Alerts before stockouts occur

5. Human Error Prevention

Health & Safety Integration: Profile system tracks technician health conditions and flags inappropriate task assignments
3D Models of Servers: Technicians can view 3D models of the server shelves and click on hotspots on the server containing common concerns
Task Difficulty Evaluation: Warns technicians about physically demanding assignments
AI-Powered Device Assistant (/field-ops): Context-aware equipment guidance using Google Gemini 2.0 Flash API
- 10+ device database (NVIDIA H100, QSFP-DD 400G, NVSwitch, etc.)
- Predefined and custom questions
- Brief, actionable responses (max 300 tokens)
Visual Cabling Health: Bundle health visualization to prevent installation errors

How we built it

Tech Stack

Frontend & Styling: Next.js 16.0.1 with React 19.2.0, TypeScript, and Tailwind CSS 4.0 with custom dark theme
Backend & Database: Next.js API routes with server-side rendering and MongoDB Atlas for scalable, global data replication
AI Integration: Google Gemini 2.0 Flash API for device assistant and ticket parsing, Gmail API for email ingestion
State Management: React Context API with localStorage persistence
Algorithms: Custom pathfinding with mathematical optimization

Architecture

We built a fully integrated ecosystem where every module communicates:

Inventory system feeds into work order generation
Health profiles inform task routing and warnings
Communication hub auto-updates tickets across all systems
AI assistant provides context-aware equipment guidance
Floor plans visualize optimized technician routes

Mathematical Models

Predictive Depletion Calculation:

$$\text{Weeks Remaining} = \frac{\text{Current Quantity} - \text{Threshold}}{\text{Adjusted Weekly Usage}}$$

Where adjusted usage accounts for trend:

$$\text{Adjusted Usage} = \text{Avg Weekly Usage} \times \left(1 + \frac{\text{Usage Trend}}{n \times 10}\right)$$

Depletion Status Classification:

$$\text{Status} = \begin{cases} \text{Depleted} & \text{if } w \leq 0 \ \text{Critical} & \text{if } 0 < w \leq 1 \ \text{Warning} & \text{if } 1 < w \leq 2 \ \text{Healthy} & \text{if } w > 2 \end{cases}$$

where w = weeks remaining

Key Design Principles

Health & Safety First: Technician well-being integrated into every task assignment
Dark Theme UI: Consistent cyan/blue aesthetic for data center environments
Mathematical Determinism: Explainable algorithms, not black-box AI
Real-time Sync: Live data across all modules
Smooth Animations: Fade-in, slide-in transitions for polished UX

Challenges we ran into

Multi-Constraint Pathfinding: Balancing distance, elevation, priority, temperature zones, AND health conditions required extensive algorithm iteration and weight tuning.
AI Integration Complexity: Integrating Google Gemini API while maintaining brief, actionable responses (300 token limit) required careful prompt engineering.
Real-time Multi-System Sync: Ensuring inventory, work orders, health profiles, and communication all stayed synchronized without lag required careful MongoDB schema design and efficient API routing.
Health Data Sensitivity: Building a health tracking system that's useful for task assignment while respecting privacy and avoiding discrimination required thoughtful UI/UX design.
Email/Slack Parsing: Extracting structured ticket data from unstructured communications with varying formats required robust AI parsing logic.
UI Complexity Balance: Creating an interface powerful enough for complex data center operations but simple enough for technicians in high-pressure situations.

Accomplishments that we're proud of

Complete Ecosystem Integration: All five NMC challenge areas work together seamlessly in one platform
Health-First Design: Unique integration of technician health data prevents injuries and ensures appropriate task assignment
Production-Ready Code: Clean TypeScript with proper error handling, type safety, and responsive design
AI-Powered Intelligence:
- Device assistant with 10+ equipment database
- Automatic email/Slack ticket parsing
- Context-aware responses optimized for field work
Mathematical Optimization: Novel pathfinding algorithm considering factors rarely seen together (elevation + zones + health + priority)
Predictive Analytics: Trend-based inventory forecasting prevents stockouts
Zero Manual Entry: Fully automated ticket creation from multiple channels

What we learned

Human-Centered Automation: The most powerful automation still requires human well-being at its core
System Thinking: Interconnected systems multiply their value when properly integrated—inventory informs routing, health profiles inform assignments
Algorithm Design: Balancing five constraints (distance, elevation, priority, zones, health) in optimization requires careful mathematical weighting
AI Prompt Engineering: Crafting prompts that produce brief, actionable responses (300 tokens) while maintaining context and accuracy
Next.js Architecture: Deep understanding of server vs. client components, API routes, SSR, and real-time data patterns
MongoDB Best Practices: Global client management, connection pooling, schema design for real-time sync
Health Data Design: Building useful health tracking that respects privacy and avoids bias
Real-World Impact: Designing for actual technician workflows in $2.8B facilities, not theoretical scenarios

What's next for DC Daddy

Immediate Enhancements

Mobile App: Native iOS/Android app for on-the-go access
Voice Commands: Hands-free operation for device assistant
QR/Barcode Scanning: IoT sensor integration for real-time equipment monitoring
Advanced Analytics: Machine learning for predictive maintenance

Goal: Deploy to all NMC facilities globally, preventing downtime and saving millions while keeping technicians safe and healthy.

Business Impact

Significant Cost Reduction across all facilities through automation and prevention
Enhanced Uptime and Reduced Emergency Procurement through predictive inventory management, accurate forecasting, and faster incident response
Increased Technician Productivity via optimized routing and reduced manual work
Improved Worker Safety by preventing injuries through health-aware task assignment
Smarter Operations for massive compute infrastructure at a global scale

Technical Novelty

Integrated Intelligence Layer

Unified data ecosystem connecting work orders, inventory, routing, and health profiles
Real-time + predictive analytics fusion
Multi-channel communication ingestion

Algorithm-Driven Optimization

Custom pathfinding with 5-constraint optimization
Mathematical clustering (deterministic, explainable)
Trend-based forecasting without black-box ML

Health-First Design

Proactive task evaluation based on physical requirements
Automatic warnings prevent injuries
Context-aware assignment system

Production-Ready Architecture

Globally scalable with distributed data replication
High-performance API with minimal latency
Type-safe development with modern responsive UI

Project Status

Production Ready: All core features implemented, tested, and fully integrated (tickets, inventory, routing, health, AI)
AI Assistant: Fully functional with Gemini 2.0 Flash API for device guidance and ticket parsing
Health Integration: Profile system with automatic task warnings for worker safety
UI/UX: Polished dark theme with smooth animations, responsive across all screen sizes