Natalie Glance’s Post

Our first-ever Duolingo Engineering Summit was a huge success 🎉 We filled the PPG Paints Arena in Pittsburgh and had a full day of team bonding and learning. We kicked it off with a keynote that went through the “eras” of Duolingo Engineering, and then Duos could choose a “track” to follow for the day, including: iOS, Android, AI, Backend, Platform, and Career Development. Each track consisted of talks from fellow Duos and training on relevant topics (i.e., on the Android track, you might attend “Building & Scaling Android Math” and on the Platform track, you might attend a talk on Kubernetes). I got to attend several sessions, including: “How to Debug Anything,” “Career Development for Senior+”, “Allyship for Managers,” and “How teaching Math is different than language.” I had a lot of FOMO for the sessions I didn’t attend! It wasn’t all work: we had a happy hour, game time, and our in-house design team helped us commemorate the event with some exclusive swag 👟 We’ve grown so much in the last decade, from nine engineers to a team of nearly 400, and this offered the opportunity for almost the entire team to come together, share knowledge, and build connection. Thanks especially to Brian Dewey, who first pitched the idea to me and Severin Hacker, and Cara Silverman for helping bring this idea to life! In many ways, it was the best conference I’ve ever been to 😎 #engineering #engineeringleadership #teambonding

Did you catch the Duolingo feature in the The Wall Street Journal? I love this quote from Ben Cohen: “I wrote an entire book about the power of streaks, but I’ve never come across an example of streaks as powerful as Duolingo.” My husband was looking over my shoulder as I read the article, which cited the factoid that the longest active Duolingo streak is now longer than 4,100 days (that’s over 10 years!) and that more than 70% of the app’s more than 30 million daily active users have weeklong streaks. He asked me what my streak was (it’s 2,670 days) and then asked me how many Duolingo users have a streak as long as mine. So I looked it up (I love that our metrics tools are so good that I could find the answer in seconds): 28,618 people have a streak as long or longer than mine! How did our streak become so powerful? It’s the result of more than a decade of a/b testing (which I’ve talked about before) to make something that’s otherwise difficult (learning a language) into a daily habit that is fun and engaging. And we’re continuing to build new features that will make you want to return to your lessons day after day. Stay tuned to find out what those features are 😉 https://lnkd.in/emaXwPrU

Duolingo engineers get to take the lead on solving unique technical challenges they might not encounter anywhere else. These 3 problems are just a sample of the exciting work happening across the org. And what I love about these is that they are so clearly tied to our mission, and the solutions help serve learners all over the globe 🌎 ❓ How do we ensure all of our learners have an equally high-quality experience? ❓ How do we scale personalization? ❓ How do we make English certification testing more accessible? Get a peek at the creative engineering work underway at Duolingo, and come join us if this work seems exciting to you! ➡️ https://lnkd.in/eSJYjYPG https://lnkd.in/eUjkJrvj #engineering #engineeringjobs #engineeringsolutions

Last month, engineering leads at Duolingo held our first “engineerication.” (I got the idea from Stripe.) Basically, we took a week off from our regular jobs to sit together and code in order to see what our engineers do daily, get back into the tech stack to see what can be improved. 👩💻 Turns out, a lot has changed in the years since I last wrote a line of code 😅Some things have gotten much easier. Codespaces, for example, allows engineers to easily create a coding environment in the cloud instead of on their local machines. And GitHub Copilot auto-suggests chunks of code. But there were 3 big takeaways I had for how we can improve our engineers’ experiences: 1. On the first day, I set out to do a few small code changes. But as a new engineer, you can’t just sit down and make a change immediately, even a tiny one. You have to set up your environment and figure out where the code needs to go. Then you have to build the code, test it, get it reviewed, and push it live to production. All of this takes time—sometimes, too much time. I want to make it easy for new software engineers to be productive right away without having to set up their development environment or having to learn a lot of arcane tools, processes, and magical incantations. 2. When I started at Duolingo nine years ago, we had a few GitHub repositories. Now we have dozens of microservices. This means that even a small code change may require updates to several repositories. If you’re not regularly working in a given repo, this is much harder, especially if it’s one that your team doesn't own. More standardization is needed to make it easier to code across our various microservices and client repos (we’re starting to do this, but more can be done!) 3. Code review turnaround time slowed us down. We averaged about a 24-hour turnaround, but when making atomic (easily revertable) pull requests, lots of PRs stack up and eventually lead to merge conflicts, etc. In practice, I could see engineers cutting their turnaround time by making PRs that are less atomic, which leads to messier reverts and a harder time diagnosing bugs. Whenever software engineers are tempted to take shortcuts because the current process is too slow or painful, that’s an opportunity for improvement. Overall, taking a week to go hands-on with the code is a great way to build better processes and teams. It was an incredibly valuable week that helped us visualize the steps we can take to remove pain points and speed up some unnecessarily slow processes. 🐢 (One more learning: Next time, I’m going to invite one of our software engineers to pair program with me 😅) Help us develop new products and processes! We're hiring: https://lnkd.in/eSJYjYPG #engineering #processimprovement #leadershipidea

Duolingo’s Q1 results are in, and 2024 is off to a strong start. One of the things I’m most excited about is our effort to better serve a global audience of English learners. The launch of our standalone intermediate-level English course is a big step towards delivering on our mission to develop the best education in the world and make it universally available. In case you missed it: our latest efficacy research showed that learning English on Duolingo is already quite effective. Our continued investment will only improve this experience! Read more from Luis von Ahn in his letter to shareholders: https://lnkd.in/dKk_Pf48 #earningsreport #quarterlyresults

I'm an engineer... of course I ran an errant crawler that triggered Google to cut off our entire office.... (true story from years ago!)

The Growth Area at Duolingo is responsible for iterating on and testing the best ways to drive... well, growth! Here are a few projects they work on: 🔔 Notifications 🔥 Streak mechanics 🏆 Leaderboards 👯♀️ Friends Quests 🎆 Our annual Year-in-Review That's just the tip of the iceberg. We're hiring for an Engineering Director who will work with the Growth Area, and help innovate and drive new strategies to grow our learner base! 📈 If these projects sound fun to you... come join us! https://lnkd.in/dKb_Jngv #hiring #engineeringjobs

Everyone is talking about LLMs and generative AI these days—what I get really excited about is all the possibilities of combining these with other cutting-edge technologies. Here are two examples: 1️⃣ Birdbrain is Duolingo’s machine learning model for predicting whether a learner will get a given exercise right. Using data from the billion exercises our users do per day, Birdbrain can generate practice sessions that land exactly 🎯 in your sweet spot: not so easy that you don’t learn anything new, but not so challenging that you get discouraged. (Read about it here: https://shorturl.at/lLM38) Right now, our session generator tailors personalizes lessons from a large set of existing exercises. It’s an extremely effective way to learn—but like all practice, it can get a bit repetitive. Here’s where generative AI comes in: instead of picking from existing exercises, practice sessions can be built from a much larger pool of sentences built with the help of generative AI. These sessions still help you practice concepts that you’ve already learned, calibrated to your exact level, but the content itself will be new to you! Exercises start to feel less like practice and more like a real-life situation of speaking in a new language, where you won’t understand everything and will have to figure things out from context. 2️⃣ I’m also excited about our collaboration with Rive, an animation platform that uses state machines to describe how animated objects move. Our animators create a library of animations by hand from “rigs” created in Rive, that can be manipulated (like puppets!). Our engineers can then use genAI to write scripts that can trigger playback of animations, change the position of assets, move the camera, and so much more! There’s a world of possibilities for the interactive usage of the animations we create! (The character’s mouth movements, aka “visemes”, are a great example of this: https://shorturl.at/AEKX7). This allows our animators to do much more, much more quickly—and there’s a lot of potential for future features combining this technology with the creative capabilities of generative AI. 👀 Stay tuned to see the exciting developments we’re going to discover next! 😎 And join us if this sounds exciting to you: https://lnkd.in/e9c8yXRz #duolingo #languagelearning #engineering #innovation #generativeai

Today's recommended reading: a post all about Duolingo's growth from Sean Colombo, our VP of Core Product Engineering. He includes 6 guidelines we use that ensure we’re moving quickly and striving for growth without sacrificing quality, and they could apply to any org: https://lnkd.in/eCExZXNB