Cailyn Hansen

Duolingo

I started at Duolingo in Dec of 2021. I spent nearly 4 years there across 4 teams: Duolingo for Schools, Media Learning, Growth Infrastructure, and Notifications. I recently left (Nov 2025) to take a career break and explore more of my passions.

Duolingo for Schools

TBD

Media Learning

I was the tech lead of Media Learning and architected and implemented the backend for DuoRadio. I supported a team of 7 engineers across all platforms. We also built out the initial version of Duolingo Adventures and prototyped the first scaling efforts for DuoRadio, which eventually led to a 10x increase in content

Growth Infrastructure + Notifications

I was the engineering team lead for both of these teams. I led Growth Infrastructure (GINFRA) from July 2024 to the start of 2025 at which time we transitioned into a more product-focused team building out and maturing our Notifications system and experience.

While I led GINFRA, we owned 27 backend services ranging from the streak, XP calculation, notifications, and rewards/achievements. I managed a corresponding annual budget of $2 million in cloud costs, which we repeatedly reduced costs quarter over quarter.

On Notifications, for the first 9 months, I wore both engineering and product hats, leading 4 to 5 engineers and setting strategic direction. We shipped over 100 experiments contributing over 25% of DAUs from launched experiments in the Growth Pillar. In September, we gained our long-awaited product manager partner and dedicated data science and machine learning support.

My team reduced number of users getting substantially delayed notifications by 89% in the course of our first 6 months and undelivered notifications by 7M/day.

Two of my favorite experiments we launched were:
1. Delaying notifications that would be sent too close together. By ensuring non-urgent notifications didn't send within 30 minutes of each other, we not only improved the user experience but also substantially increased DAUs and reduced total notification volume.
2. Drop certain notifications after streak extension. By limiting how much we reach out to users after they have already been active on the app, we added a record number of DAUs in a single experiment and reduced notification volume by over 11M/day.

We also undertook a migration of our underlying reminders database, which was experiencing repeated stress and leading to outages, saving a 6-figure sum of money in cloud costs and increasing our reliability runway by over 5 years.

Miscellaneous

I also served as the PagerDuty rotation lead for the Language Learning organization. During my time, I developed a shadowing system for new engineers to get familiar with incident command and reliability improvements. This model was eventually adopted across the company. I also worked with tech leads across the organization to substantially reduce noisy alarms that a) did not impact the user experience, or b) did not require intervention.