Predict equipment failures before they happen.

An all-in-one, non-invasive, self-charging device that detects problems in HVAC units before they turn into catastrophic failures.

Talos Dashboard

Live anomaly trend

Asset

Walk-in cooler

Attention

12:00Now

Active alerts

Loud knock detected

Walk-in cooler compressor

Critical

Unusual vibration pattern

Back room refrigeration

Warning

Compressor balance drift

Rooftop HVAC unit

Watch

Talos AI

What's wrong

Talos detected a sudden shock event.

Why it might have happened

Possible causes include:
– Equipment being hit or dropped
– Loose parts due to wear and tear
– Improper installation or mounting

How to fix it

Inspect the equipment for any visible damage or loose parts. Tighten any loose bolts or connections.
If no issues are found, or if damage is evident, call a technician immediately.

Ask a follow-up question...

Impact at a Glance

Unplanned faults are expensive, and often preventable.

$80 Billion+

Annual U.S. loss from unexpected HVAC breakdowns

$8k-$12k

Cost per unexpected breakdown

1.2 Million

Small business locations in scope

$30B → $100B

5-year market growth projection

Loading exploded assembly

Preparing grouped CAD layers for a scroll-driven teardown.

Exploded hardware view

Our Product

Easy-to-use hardware

Our sensors are completely non-invasive and retrofittable — no technical experience required. Simply set and forget, as they are entirely self-powered and magnetize to any existing HVAC unit.

Easy, non-technical setup for customers

Retrofittable to existing HVAC systems

Non-invasive installation with no equipment changes

Magnetically attaches directly to the unit

Alerts communicated through mobile app

Users receive detailed, actionable insights explaining what happened, why the fault occurred, and exactly how to fix it. Our AI chatbot — available 24/7 — makes it simple for non-technical owners to use Talos and ask follow-up questions at any time.

Our Solution

How It Works

Easy installation

Attach device to HVAC within seconds using its magnets — no technical expertise needed.

30 sec setup

Detect vibrational patterns

Hardware continuously captures vibrational data and streams them to our cloud server.

Continuous sensing

AI diagnosis

Our model cross-references vibration data against failure signatures to detect issues weeks before they occur.

Now+2 weeks

Failure

△ Alert issued here

Weeks of notice

Notify users

You receive a push notification with what's wrong, why it's happening, and exactly how to fix it.

HVAC Alert

Compressor bearing wear detected. Service recommended within 10 days.

Mobile + SMS

Our Founders

Amitav Kohli

Carnegie Mellon University

Electrical Engineering

Vibration analysis experience at NASA and ACTIVECharge. Predictive analytics and engineering experience at Constellation Energy.

Kevin Hwang

Johns Hopkins University

Computer Science

Time-series research at JHU and quantitative analyst experience at IAE.

Himani Musku