Below is a summary of my professional experience.
At Duolingo, I primarily implement and maintain internal tools that improve quality of life for designers, creative technologists, animators, and other engineers.
I'm also an active Android engineer, developing features for the Android app, rolling them out as A/B experiments, and on occassion managing the Android release.
My stack primarily consists of Python when working on backend services and Kotlin when working on the Android client. The complexity of my projects, however, often require me to be quick on my feet and learn many new tools and frameworks (eg: Alembic and SQL, AWS and S3, Sentry, etc).
As an intern at Duolingo, I built an internal tool to allow assessment scientists on our Duolingo English Test (DET) team to run studies and gather user data from optional practice test flows.
My tool allowed the scientists to collect and store mic, screen, keystroke, and audio transcription data from hundreds of test sessions involving users who opted in to the data collection.
I then collaborated closely with the scientists and linguists to analyze the collected data and discuss how we could use it to refine the ability of certain question types to measure English proficiency.
Throughout my junior and senior years at Columbia, I served as Peer Leader for the Emerging Scholars Program, a small one-credit workshop class designed to introduce students to broad Computer Science topics such as algorithms, cryptography, computer vision, Natural Language Processing, Machine Learning, and computer ethics.
As Peer Leader I put my professor hat on and gave lectures, facilitated discussions among my students, answered their questions, and connected them to campus resources. I even had my own TA (we called them workshop assistants!) helping with classroom discussions.
During the summer of my sophomore year, I had the blessing to go back home to Addis Ababa and work remotely for an edtech startup dedicated to providing educational resources to Ethiopian students.
Alongside three intern peers, we developed and deployed a Typescript-based Telegram bot that periodically sent English grammar lessons to a Telegram group of more than 6000 Ethiopian high school students. For this we used Telegram's Bot API to schedule and send lessons, as well as tweak them based on feedback.
Our MVP used a database of english lessons that we scraped from English text books, but our final version complemented these with lessons generated on the fly using the OpenAI GPT-3 API.