Building Better Speech Technology

Our goal is to build speech technology that’s more inclusive and fair, for everyone.

Today, most systems and models are trained and measured with a tiny selection of available voices that doesn’t represent the world.

We build models, benchmarks, apps, datasets, and other tools that help close the gap. We love collaborating with researchers who align with our goal of better, equitable speech technology.

A selfie of the Koel Labs founders, Aruna Srivastava, Ruslan Mukhamedvaleev, and Alexander Metzger, holding their badges at Interspeech 2026.

The Team

Our Co-Founders

Koel Labs began with three students sharing a common vision. As immigrants and children of immigrants, the experience of language barriers shaped our understanding of how pronunciation challenges impact confidence and opportunities.

Work With Us

Let's Build It Together

We partner with researchers, clinicians, developers, and others on applied speech research.