1:44 - Falling in love with machine learning
15:03 - Developing a prototype at HackZurich
23:05 - Podcasts as the world’s largest knowledge base
40:43 - Doing B2C in Switzerland
49:15 - Benefiting from the Swiss startup mafia
Kevin Smith is the co-founder and CEO at Snipd, an AI-powered podcast player built to unlock the knowledge in the podcasts you listen to. Originally from England and Germany, Kevin came to Zurich to get his MA from ETH in Quantitative Finance. After a time working for banks like Julius Baier and UBS, Kevin decided to join the startup Contovista, where he worked up to becoming head of analytics and AI.
In 2020, he decided to put his love for machine learning to good use and participated in Hackzurich together with his ETH friend Ferdinand Langnickel, at a time when a breakthrough in self-supervised learning had just happened. Having originally planned to create a prototype around the topic of meeting notes, they eventually pivoted to podcasts, and ended up winning that year’s Hackzurich. In 2021, Kevin, Ferdinand and third co-founder Mikel Corcuera Lejarreta created Snipd.
The problem Snipd solves is pretty intuitive: the audio medium is great for both consuming and producing content, but not so much for interacting with it, i.e., for storing it and consulting it. It’s pretty hard to skim through a 2h podcast episode to find what you’re looking for. Through its AI algorithm, Snipd allows you to do just that: when you’re using their podcast player, whenever you listen to something which particularly interests you, you press a button in the app or you triple click your headphones, and then AI will analyze the content you just selected to determine the appropriate start and end point, provide a transcript, a summary and a title. Snipd also provides you with automatic chapters (with titles and summaries) for each podcast episode.
By providing this service, Snipd is able to collect data about which podcast moments are interesting to each user, and this data will soon allow them to make predictions on what each specific user may be interested in listening to.
"I’ve come to the realization that creating things is something which fundamentally makes humans happy."
"Something you should ask yourself before you start a startup is: Can you see yourself working on this for 10 years?"