At sixteen, Jessica Powell had a problem with the family car: its CD player did not work. Her solution involved her father’s dictaphone, tapes of Green Day, and a summer of pressing rewind. The albums were Kerplunk and 1,039/Smoothed Out Slappy Hours. The listening equipment was hardly glamorous. The attachment was durable.
Years later, Green Day came to the company she had co-founded with a problem of its own. The original master tapes for Kerplunk were lost. Could AudioShake separate the instruments and vocals from the finished recording? For the album’s thirtieth anniversary, the band wanted to give listeners something to play with.
AudioShake produced the parts. A version of “2000 Light Years Away” could then be assembled with the vocals, drums, and bass, leaving a vacancy where the guitar belonged. Fans could supply it themselves through TikTok’s Duet feature. The teenager repeatedly rewinding a tape had become a founder helping other listeners find their way into the band.
It is a useful introduction to Powell because it starts with an actual wish. Her working life has included music copyright, a senior position at Google, and a novel making fun of the industry that employed her. The through line becomes easier to hear when you begin with the person listening.
Paris before product launches
Powell studied comparative literature at Stanford, graduating in 2000. She met Luke Miner, who would become her husband and AudioShake co-founder, in their freshman dorm on the first day. Their shared history long predates the company. So does her interest in music, language, and what people make with them.
Her early career took her to CISAC in Paris, the international confederation representing authors’ societies. Music copyright places a practical question behind the pleasure of listening: who made this, who controls its use, and how do they get paid? It was an early encounter with the interests of creators that would return in a different form at AudioShake.
She also wrote a book. Published in 2006, Literary Paris: A Guide follows thirty writers through the places associated with their lives and work, from Molière to James Baldwin. More than a hundred sites turn literary history into a walk through a city. Long before she was explaining software, Powell was connecting people’s creations to the circumstances that made them possible.
Her biography has room for translation, journalism, editing, and a stint as a wild boar chaser. It is an unusually lively list of previous employment. Few executive introductions invite a follow-up question about livestock before they reach the management team.
The view from the communications desk
Google brought another set of countries and responsibilities. Powell worked in London, Tokyo, and eventually the Bay Area. She became vice president of communications and served on the company’s management team, reporting to its CEO. Her career also included a period as chief marketing officer at Badoo.
Communications at that level involves decisions about products, policies, and the arguments surrounding them. Powell has described the work as heavily strategic. Someone has to understand what a company is doing before deciding how to explain it. That position can offer a close view of the machinery and an equally close view of the gap between intention and reception.
Working in Europe mattered to her understanding of technology. Questions about products and their implications arrived from people outside the industry’s familiar assumptions. Defending a new service required thinking through why somebody else might object to it. The product’s existence did not settle the argument.
There is something fitting about a comparative literature graduate doing this work. Different audiences can encounter the same thing and come away with different meanings. A communications department may be expected to smooth out those differences. A novelist can make very good use of them.
A company with a fictional address
Powell’s novel, The Big Disruption: A Totally Fictional but Essentially True Silicon Valley Story, concerns Anahata, a technology company with a prodigious appetite for its own mythology. Among its inhabitants are an exiled prince working in product management, warring professional tribes, and female employees whose ambitions receive considerably less attention than the ambitions around them.
The comedy depends on recognizable workplace behavior enlarged until its absurdity becomes difficult to ignore. Authority attracts agreement. Fashionable language can pass for understanding. The corporate promise grows so large that the people inside it start to look like inconvenient details.
Medium published the novel digitally in 2018, with a print edition following in 2019. Powell directed proceeds from the Medium collaboration to #YesWeCode and Book Trust, supporting access to technology careers and books respectively. The book’s author biography described someone who could be both fascinated and unsettled by the industry.
That combination has lasted. Writing satire did not send Powell into permanent exile from technology. She continued building a company. The novel and the startup sit alongside each other, which makes her a more interesting observer than a simple conversion story would allow. She can find a practice ridiculous and still care about what the people doing it might produce.
“I hate introducing myself, so this will be very short.”
Jessica Powell
Her public humor often lands on herself first. When a podcast host announced that he had bought her book, she responded, “That will pay for my retirement, I’m sure.” For someone who spent years overseeing corporate language, she is willing to puncture the ceremony of an introduction.
The karaoke question
The idea behind AudioShake grew partly from life in Tokyo. Karaoke offered a simple frustration: what if you wanted to sing a song outside the available selection? Turning an existing recording into an instrumental would let the catalog expand beyond the choices already prepared for you.
Powell and Miner returned to the question as hobbyist musicians. Could software pull the voice away from the rest of a song? Could it isolate the bass line someone wanted to learn? At first, the result looked like an entertaining experiment. Music industry contacts began pointing out uses with more immediate commercial importance.
A music supervisor might need an instrumental to fit beneath a scene. A rights owner might want parts for a remix or an immersive mix. Those conversations gave the founders a clearer picture of the work. Powell has described speaking to sync teams and finding interest quickly. The customer’s problem gave the demonstration a destination.
Conceptual illustration. Separate parts give an editor more choices.
AudioShake was founded in 2020 and launched publicly in 2021. That August, it announced a win in Sony’s Demixing Challenge, where entrants tested their models against the same recordings. The contest supplied an external comparison for a business whose demonstration depends on an intimate judgment: what does the result sound like?
The karaoke idea had opened into something broader. A finished mix is convenient for listening but restrictive when someone needs to change one element. Original session files may be missing, unavailable, or never have existed as separate tracks. Source separation gives the person working with that recording another route.
The recording gets another job
Powell’s account of the business repeatedly returns to the owners of existing work. A song may have another life in a film, a different mix, or a fan’s performance. The usefulness of the technology depends partly on those relationships. A clean isolated vocal is a technical result; permission and purpose determine what happens next.
AudioShake’s recognition widened, including a place on TIME’s Best Inventions list in 2023. In 2024, the company joined the Disney Accelerator and announced a collaboration with Disney Music Group. Its applications also expanded into separating film and television audio into dialogue, music, and effects, giving production teams parts they could work on independently.
Consider dubbing. A team preparing a program in another language needs room for replacement dialogue while preserving the surrounding sound. Separating the existing mix can help that work proceed. The connection to karaoke is surprisingly direct: both require making space inside something already recorded, though the people doing the work have very different deadlines.
In October 2025, AudioShake announced a $14 million Series A led by Shine Capital. At the time, the company reported more than forty enterprise contracts and more than one hundred million minutes of audio processed. Those are company figures tied to that announcement, rather than a running measure of its current size.
The customer examples had grown beyond a music startup’s first pitch. They included film production, sports footage, and structured audio for machine training. Yet the basic operation remained recognizable. Start with a recording in which several things happen together. Give someone access to the parts.
Making room for another voice
By September 2026, the problem included people talking over each other. AudioShake’s Multi-Speaker 2.0 release was designed to separate overlapping voices into individual tracks. It also supplied confidence markers directing editors toward moments that needed human review.
The appeal is easy to understand if you have ever listened to an animated conversation. Laughter, interruptions, and someone finishing another person’s sentence can be the parts worth keeping. They can also be the parts that make an editor wish everyone had waited politely for a turn. Powell’s explanation of the release emphasized preserving those performances while giving editors control over the audio.
Her broader ambition is for people to interact with sound as readily as they do with images and video. In her Webby Awards interview, she also put a condition on useful technology: “No one is served if we in tech just build the things that we think people need or want.”
For Powell, the question keeps coming back to another person’s use for the work. The singer wants a backing track. The editor needs the dialogue. The teenager wants to play along. There is room in that last wish for quite a serious business, and for the pleasure that made anyone press rewind in the first place.