Democratizing
Diagnostics
The automotive repair industry has operated as a black box for decades. The asymmetry of information between mechanics and drivers has created an environment where diagnostic fees are inflated and consumers lack the data necessary to verify repair quotes.
Carithm was built to eliminate that asymmetry using acoustic artificial intelligence. Every mechanical fault—from a failing water pump to a degraded wheel bearing—emits a distinct acoustic frequency. By processing these sound waves, we deliver dealership-level diagnostics directly to the consumer.
From Enterprise to Consumer
Carithm initially began as an enterprise B2B solution. Early iterations of our predictive AI were pitched in boardrooms to service centers across the UAE, designed to forecast part wear-and-tear under extreme Middle Eastern climatic conditions.
In August, we executed a strategic pivot to a consumer-facing model. Rather than locking our diagnostic architecture behind enterprise contracts, we deployed it as a free, universally accessible web application.
Data & Architecture
Training an accurate acoustic model requires verified data at scale. We engineered automated pipelines to process thousands of documented mechanic diagnostic recordings, mapping the precise frequencies of mechanical faults to build a proprietary dataset entirely in-house.
Today, Carithm processes audio for over 10,000 users a month. By optimizing upload constraints and maximizing edge computing, operating overhead remains negligible—allowing us to maintain the tool as a free public utility while the business scales.
I began developing the core architecture for Carithm at 17, combining a lifelong study of motorsports with a drive to solve systemic inefficiencies in the automotive market. Today, as a 19-year-old mechanical engineering student originally from India—now living, studying, and running the company from the UAE—I am proud to see the platform empowering tens of thousands of drivers globally.
