AudioShake

Powell brings a wealth of experience from her time at Google, where she was a member of the management team and ran global communications, and a career start at CISAC, the International Society of Authors and Composers. Miner’s background is rooted in data science and economics, with previous roles at Plaid and a Ph.D. in Economics from the London School of Economics. Their combined expertise in technology, media, and machine learning formed the foundation for AudioShake, created to solve the inefficient and laborious process of audio track separation.

The company's core service utilizes deep learning models to isolate vocals, instruments (like drums, bass, and guitar), dialogue, music, and effects from a single mixed audio track. This capability is valuable across several industries. In music, clients include all three major label groups and prominent publishers such as Hipgnosis, Primary Wave, and Concord. They use the technology for remastering older songs, creating immersive mixes like Dolby Atmos, generating instrumentals for sync licensing, and producing samples for new tracks. For instance, the technology was used to isolate a sample of Ol' Dirty Bastard for a SZA song and to help release De La Soul's catalog on digital platforms. Beyond music, Hollywood studios, dubbing companies, and transcription services utilize the platform to clean up dialogue, separate music and effects for localization, and improve transcription accuracy.

AudioShake's business model is usage-based, with clients paying for the quantity of stems they create. Access is provided through a web-based platform for professionals and independent artists, as well as via an API and SDK for developers to integrate the technology into their own applications, including gaming, social media, and fitness apps. The company has secured over $5 million in funding, with investors including Metallica-backed Black Squirrel Partners, Crush Ventures, and Indicator Ventures.