I recently read a fascinating story about a particle physicist turned co-founder of an AI-based language learning app called Lingvist and felt compelled to find out more.
Mait Müntel was part of the Higgs boson discovery team at CERN, where he was frustrated by being unable to speak French fluently. In typical scientist fashion, he decided to look at the French language as a mathematical problem that could be fun to solve– and soon used CERN’s machine learning capabilities to run his language-as-math hypothesis, then began building out software to test it.
Within a year, he was able to study French for 200 hours, then pass a high school French test. So he left to start Lingvist, a company that uses AI to create personalized language experiences, with 1.4 million language users.
Mait joins me on Tech Talks Daily to chat about how machine learning is significantly accelerating the language learning process, and how important scientific theories have been to Lingvist’s success.
I learn how he got the idea to approach language learning like a mathematical challenge and decided to build his own solution instead. Mait’s prototype drew upon the machine learning work he had used while searching for the boson particle, along with the basics of language statistics and cognitive science theory.
After learning with a homemade prototype for 200 hours, he put himself to the test and took a high-school-level French exam, which he passed with a strong score. That was the seed from which Lingvist took root.
Mait took those principles he had used to teach himself and started developing them for anyone to use, with the goal of significantly accelerating language learning and reducing inefficiency in the learning process.
They built the proof of concept after receiving the Prototron Grant, a seed-funding initiative for budding startups, in March 2013. In 2014 Lingvist was selected from a pool of over 1,500 applicants to participate in the world-renowned TechStars London accelerator program.
Today, Lingvist is an eclectic team of more than 40 languages and technology lovers hailing from around the world. Members include alumni from Skype and CERN, and specialists in areas ranging from AI and computational linguistics to UX and product development.
Subscribe to the Tech Talks Daily Podcast