Enterprise Architecture.
On-Prem Control.

Proven stack. Modular architecture. Security at every layer. A•IQ deploys inside your network, under your control, with zero cloud dependency.

System Architecture

Resilient by Design

Every component runs independently. APIs are isolated. Services can be updated, scaled, or replaced without taking down the system.
Isolated Microservices - Dedicated process for every function (Alarms, PPCL, ML). Failures are contained, never systemic.
‍Native Windows Services - Zero 'script fatigue.' Deploys as a managed, auto-restarting background service. No open terminals required.
‍Nginx Reverse Proxy - Traffic routing, load balancing, SSL termination.
‍API-First Integration - Full data accessibility. If A•IQ can see it, your other systems can query it via standard REST endpoints.

Database

Time-Series at Scale

Optimized for the high velocity and volume of modern building automation data.

PostgreSQL

Battle-tested relational database

TimescaleDB

Time-series extension optimized for high-volume sensor and alarm data

Garnett

Next-gen, Redis-compatible caching. Ultra-low latency with enterprise-grade security.
Millions of records | Sub-second queries | No cloud dependency

Security First

Security Isn't a Feature. It's the Foundation.

A•IQ was designed for environments where security isn't optional - federal facilities, healthcare, financial institutions. Every layer is locked down.

Authentication

FusionAuth - Enterprise identity management
Role-based access control
SSO-ready architecture

Secrets Management

OpenBao - Secure secrets storage
No hardcoded credentials
Encrypted at rest and in transit

Transport Security

SSL/TLS everywhere - API to API, client to server, service to database
Certificate-based authentication
Nginx SSL termination

Deployment

On-premises - Your data never leaves your network
No cloud dependency
Air-gap capable

ML STACK

Local ML. Full Control.

A•IQ's machine learning runs locally on your infrastructure. No data sent to external APIs. No black-box cloud services. Trained on your data, for your buildings.
  • The Stack:
  • PyTorch - Flexible model development and research
  • scikit-learn - Standard efficient algorithms for clustering and anomalies
  • TensorFlow - Deep learning and neural networks
The Methods:
Anomaly detection (Isolation Forest)
Alarm clustering (K-Means, DBSCAN)
Temporal pattern analysis (LSTM/GRU)
Severity classification (Random Foreset)
Predictive load forecasting (Linear Regression)

FRONTEND

Fluid Analytics. Instant Insight.

Aggregating historical data requires a frontend that can handle weight. A•IQ uses React’s efficient rendering to manage heavy datasets without the drag. Zoom, filter, and pivot through millions of records instantly, delivering deep-dive trends and analytics tools that feel like a native app, not a webpage.