Dossier.

There was a man in suburban Detroit who could name a car's pending repair by the sound it made coming up the driveway.

That was Evelyn Kowalski's father. The garage was small and independent and had been in the family for two generations. The man's ear was the kind of thing that took forty years to build and was very nearly impossible to teach. Customers thought of it as a trick. He thought of it as listening carefully.

Years later, after a career building acoustic models in adjacent industries, Evelyn returned to the question she had grown up around. If a great mechanic could hear what was wrong, then the patterns had to be learnable. The only real question was whether enough labeled data could be gathered to teach a model to recognize them at a useful level of accuracy. The answer turned out to be yes, but it took three years and fourteen thousand engines to find out.

Revly is led by Evelyn Kowalski, with a small team working out of Brooklyn and Detroit. Before Revly, Evelyn spent a decade in audio machine learning, most recently building acoustic event detection systems for industrial monitoring. Before that, she answered the phones at her father's shop and learned what a failing water pump sounds like before it actually fails.

My father never needed a code reader. He needed a quiet minute and his ears. Revly is for everyone who didn't grow up next to him. — Evelyn Kowalski, Founder

The team is small by design. We answer every email personally. We refuse investor pressure that would push us toward a less honest product. We are interested in being right, not impressive.