Case study 05 · Pakistan Air Force · Hyetel X · Aug 2025 – Mar 2026
Smart Building Automation System
A centralized Building Management System that monitors and controls 16+ building subsystems in real time, with AI-based anomaly detection for predictive maintenance.
85%
less downtime
The problem
Operators need one view across fire safety, CCTV, sensors, alarms, HVAC, access control, parking, elevators and generators — and when an incident fires, they need its exact location and the right camera feed immediately.
Architecture
- 01Integrates HVAC, CCTV, fire alarm, access control, parking, elevators, generators and energy
- 02Handles millions of requests with clustering, worker threads and distributed CPU/GPU work
- 03Real-time dashboards over WebSockets and server-sent events
- 04AI anomaly detection for predictive maintenance
Step through how a request moves through this system.
What I owned
- 01Designed the overall distributed system architecture
- 02Designed real-time incident communication and event processing
- 03Designed building-location and device-mapping workflows for 3D incident navigation
- 04Integrated camera APIs and live feeds (HLS, WebRTC, MJPEG)
- 05Implemented AI anomaly detection for predictive maintenance
- 06Worked on high-volume backend processing across services
Challenges
- 01Processing high-volume real-time events while keeping the operator UI responsive
- 02Coordinating many building subsystems through one centralized architecture
- 03Mapping physical building locations to 3D coordinates for incident navigation
- 04Switching camera feeds in real time without disrupting the operator workflow
Outcome
A centralized monitoring environment that detects incidents, locates them in an interactive 3D building model, opens the relevant live camera and tracks each incident to resolution — with backend optimization cutting downtime by 85%.
Stack
- Node.js
- Express
- Python
- React
- Next.js
- Socket.IO
- SSE
- Redis
- RabbitMQ
- Kafka
- Kubernetes