Analyzing how millions of people learn
Duolingo uses human computing to scale language-learning
This story is now published on 🦄 Babyunicorns.com 🦄 at https://www.babyunicorns.com/articles/analyzing-how-millions-of-people-learn
I asked a friend about his recent trip to Guatemala and he was quick to mention that how he brushed up on his Spanish with Duolingo, a go-to source for language-learners (and, apparently language brusher-uppers).
It seems like a lot of conversations about language-learning these days give a quiet nod to Duolingo, whose global reach —
170 million users, makes it the most downloaded app in “education”
— on both iTunes and Google Play. Given its scale, the day-to-day design and engineering challenges for the team are substantial.
What is Duolingo?
Quite simply, Duolingo is a language learning app. “Duo” (not Google Duo, btw) and “lingo” both come from Latin — duo meaning “two” and lingo, which stems from lingua, or tongue, meaning “a foreign language or local dialect”. I imagine that CEO, co-founder, and reCAPTCHA creator, Luis von Ahn, had this in mind when naming the company. Some might argue that we should have seen this coming with his PhD thesis at Carnegie Mellon being the first publication to use the term
“human computation” — referring to the methods that combine human brainpower with computers to solve problems that neither could solve alone.
Growing up in Guatemala City, Von Ahn saw how expensive it was for people to learn English. This observation fueled a project, which began at Carnegie Mellon University in 2009 and quickly turned into Duolingo, the company whose mission is to make language education accessible to everyone —
by empowering people with a free education. This, they believe, could change the world.
How do they keep it free? Crowdsourcing language duties to its millions of volunteers — “human computing”. Although this wasn’t the original plan, Luis von Ahn realized the potential for Duolingo and changed the app from being an enterprise offering to a free app for all.
Duolingo’s high valuation is a result of a small, gifted and hard-working team. It also helps to have a CEO with a PhD in Computer Science who specializes in solving problems of this magnitude; and who has a determination to use engineering to drive the company towards its lofty goal of bringing languages of the world to everyone.
Why Duolingo is a Baby Unicorn:
- Some of the brightest minds in “human computing” are democratizing language learning at scale
- Broad reach w/over 170M users
- Raised $108.3MM to date with investors such as Tim Ferriss, Drive Capital, New Enterprise Associates, and Union Square Ventures
- Strong engineering culture and commitment to diversity and inclusion: “It’s better to have a hole, than an a**hole”- Luis
What’s in the Stack?!
As you can imagine, Duolingo relies heavily on ML and NLP methods. Python, Java, Scala, C/C++, functional NLU, IR, SMT, SQL, ETLs, Redshift, Hadoop, Objective-C, Swift w/Cocoa Touch, iOS & Android SDK.
Know someone looking?
Duolingo is growing fast (world-wide) and looking for talented engineers as well as educators (PhD), research scientists, designers and product people.
Want to join the team at Duolingo?
Sign up on Leap.ai to get matched and directly referred to Baby Unicorns like Duolingo and many others!
