Mathematical Physics student uses Deepfake AI for viral smartphone app

A serious passion for physics was what drew Parshant Utam to study Mathematical Physics at the University of Waterloo. Little did he realize that the University’s entrepreneurial atmosphere would lead him to become a co-founder of viral smartphone app that is pure fun.

Parshant came to Waterloo as an international student from Pakistan, fascinated by theoretical physics. He was drawn by the presence of both the University of Waterloo’s Institute for Quantum Computing and the nearby Perimeter Institute for Theoretical Physics.

But artificial intelligence also held exhilarating opportunities. “I was seeing the startup environment heating up and people getting excited about AI,” says Parshant. “There is so much state-of-the-art research happening in AI and there is a lot to innovate.”

Parshant Utam is the co-founder of the viral smartphone app Wombo.ai.For Parshant, that innovation came in the form of co-founding an app called Wombo.ai. The Wombo.ai lip-synching app is powered by artificial intelligence and lets people turn their still images into divas who are swaying, rolling their eyes, bopping to the music and belting out a favourite pop tune. “It is really funny,” Parshant says. "People get surprised whenever they watch their WOMBO and ask: ‘How am I doing this?’”
Artificial intelligence is the technology behind what is known as “deepfakes,” which has raised ethical eyebrows because of how fake videos might be used for disinformation purposes. But Wombo has focused on simply being lighthearted fun.

The app has gone viral. Wombo may sound silly, but it is a serious business. “Right now, we are at around 50 million downloads. Every single day, we are adding 500,000 downloads,” Parshant says. “We have around 2 to 2.5 million daily active users coming to the app.”

The company has also received serious financing — nearly $6 million US in seed funding from a group of investors led by Global Founders Capital to support the team that consists of six co-founders plus technical employees who are working on improving and developing the smartphone app.
Parshant met the other Toronto-based Wombo co-founders — Ben-Zion Benkhin, Akshat Jagga, Angad Arneja, Vivek Bhakta and Paul Pavel — while on a co-op work term in Toronto. Soon after, the team found Navya Mehta, a Waterloo student studying computer science and finance to be their Founding Engineer. Coming full circle, Wombo is how hiring their own Waterloo co-op students.

Although there are other lip-synching apps, the team of co-founders realized that few are easy to use and able to be scaled up for mass audiences.

The Wombo team found solutions to the challenge of scalability

says Parshant, who is the AI lead. “You can make a Wombo in five seconds. It is really easy to use.”

There is a large song selection, ranging from the theme of the TV show Friends to Who Let The Dogs Out by Baha Men. One of the most popular is Tunak Tunak Tun by Daler Mehndi, which has had about 30 million downloads. The team is now working on improving and adding new features to the Wombo app, such as the ability to create singing duets.
Parshant says the University of Waterloo mathematical physics co-op program helped him build technical and team leadership skills.

The program is hard and it makes you tough in terms of the work ethic that you build. That prepares you [for] even harder challenges.

Parshant, who is scheduled to graduate in 2023, wants to continue working on hard problems in the field of artificial intelligence. “I want to find that intersection between what I can provide to the community and hard problems that need to be solved.”

But for now, he is enjoying working on an app with the sole purpose of making people happy.

Our job is to make people laugh.

Story Credit: Rose Simone, UW Faculty of Science