HOMECASE STUDYATLAS

Atlas

AI-powered building intelligence — turn complex BMS data into clear, actionable insight.

Role
End-to-End
Platform
Web · Cloud · OT
Stack
Next.js · Niagara · AWS
Focus
AI · Operations
SCROLL

01 — BRIEF

Buildings generate the data. Operators need understanding.

Traditional BMS interfaces are built around point trees and engineering workflows — not answering “what needs attention?” across a portfolio.

Thousands of HVAC points, alarms, setpoints, and schedules. Access isn’t the hard part. Clarity is. Atlas puts a modern intelligence layer on top of existing automation infrastructure.

02 — NAVIGATION MODEL

Portfolio → Building → System → Issue.

Progressive disclosure. Executives stay high. Technicians drill deep.

01

Portfolio

Instant snapshot across buildings — alarms, comfort, status, weather, and AI insights in one view.

02

Building

Dedicated operational views with floors, systems, and equipment organized the way operators think.

03

Equipment

Focused AHU, VAV, and device views — temperatures, setpoints, commands, and alarms that matter.

04

Insight

AI reads relationships between values so technicians know where to investigate first.

03 — VISION

“What needs my attention?”
Answer that first.

Not another point browser. A modern layer between Niagara infrastructure and the people running buildings every day.

04 — PRODUCT SURFACES

Built for how operators work

01

Portfolio Overview

Alarms, comfort, temperatures, equipment needing attention, weather, anomalies, and AI insights — portfolio-wide.

02

Building Intelligence

Floors, systems, and equipment in a visual hierarchy. Context raw telemetry alone can’t provide.

03

Equipment Focus

AHUs, VAVs, sensors, controllers — only the points, setpoints, commands, and alarms that matter.

04

Designed for Operators

Overview → Buildings → Systems → Equipment → Details. Everyday tasks without engineering-level navigation.

05 — AI LAYER

Not a chatbot on a dashboard — an analytical layer over the BMS.

Zone at 76.8°F, cooling setpoint 72°F, damper at 100% — Atlas can explain that the zone is above setpoint while already fully open, pointing investigation upstream.

Context fed to the model

  • Equipment points & live values
  • Alarm states & priorities
  • Setpoints & schedules
  • System relationships
  • Equipment configuration
  • Historical operational context

06 — OT ↔ CLOUD

Niagara stays in control. Atlas sits above it.

Custom Java modules form a controlled bridge — translating equipment, points, alarms, and schedules into a format the cloud platform understands without exposing controllers to the public internet.

Building Automation → Niagara → Java Modules → Cloud → Atlas

07 — SYSTEM

Full-stack cloud platform on industrial roots.

Next.js + React

01

Web UI

Responsive operator app across desktop, tablet, and mobile — shared design language for dashboards, alarms, and equipment.

PostgreSQL

02

Data model

Organizations, buildings, systems, equipment, devices, users, and operational metadata in a relational hierarchy.

Java / Niagara

03

OT integration

Custom Niagara 4 modules expose building data while keeping the operational network separated from the cloud.

AWS

04

Cloud infra

Centralized building intelligence with secure remote access — no direct public exposure of BMS interfaces.

AI Layer

05

LLM context

Summaries, anomaly surfacing, and troubleshooting assistance grounded in real equipment relationships.

Security Boundary

06

OT mindset

Controlled integration edge between building systems and cloud services — remote intelligence, isolated control.

08 — ENGINEERING

Hard problems. Clear surface.

  1. 01

    Bridging OT and cloud

    Translate points, devices, schedules, and alarms into structures a SaaS app can consume — without fighting how Niagara actually works.

  2. 02

    Modeling buildings

    Organizations → buildings → systems → equipment → thousands of points. Hierarchical, queryable, navigable.

  3. 03

    Making AI understand a building

    An LLM isn’t useful on isolated sensor values. Context — relationships, setpoints, alarms, schedules — is the real product.

  4. 04

    Deciding what not to show

    Traditional BMS exposes everything. Atlas prioritizes operational importance and reveals depth only when needed.

09 — INTERFACE

Clean. Fast. Operator-first.

Dashboard · Alarms · HVAC · Schedules · Devices. Status loud. Noise quiet.

Atlas portfolio dashboard with building health overview
Portfolio dashboard — what needs attention
Atlas alarms view with priority and acknowledgment workflow
Alarms — real-time monitor & respond

10 — ROLE & OUTCOME

Extend the BMS. Don’t replace it.

Designed and engineered end-to-end — from Niagara modules and AWS to operator UX and AI context. A bridge between decades of building automation and modern cloud software.

Niagara keeps reliable control. Atlas adds cloud connectivity, clearer visualization, simpler navigation, and AI-powered intelligence.

  • Product architecture
  • UX & interface design
  • Next.js / React
  • Backend architecture
  • PostgreSQL modeling
  • AWS infrastructure
  • Niagara integration
  • Java Niagara modules
  • Building data modeling
  • Equipment visualization
  • Alarm & schedule UIs
  • AI context & tools
  • Security architecture
  • Remote-access design

Data → Insight

LET'S WORK TOGETHER

Have something great in mind? Let's build it.

I'm always open to discussing new ideas, challenging problems, and impactful projects. Let's create something exceptional.