Summer Intern on the Home Timelines team at Twitter. I implemented improvements to the organization of tweets on the timeline through rigorous A/B testing and data-driven development of features. This involved working through the entire product feature workflow, from conception to staging and testing/experimenting, to final launch. I staged Mesos instances with Aurora and tracked metrics by directing traffic to shards with my changes. Furthermore, I wrote YAML and Ruby files to aid deployment in testing and production environments. In addition, I launched a module using Scala that aggregates tweets by hashtag, personalizes these groups to the users' networks, and then injects them into the timeline. My changes over the internship generated 10M+ user active minutes and significant increases in tweet engagement and user retention.

I teach students problem-solving strategies for computer science and discrete mathematics. I also work with students both one-on-one and in larger walk-in settings. I am a nationally certified tutor by the College Reading and Learning Association.

I created a Python program to generate 18 features that measure robust learning by a student while taking on online genetics tutor. This program parses through data given by the intelligent cognitive tutor and performs calculations to generate the features. I also wrote a preprocessing program that removes noise and prepares the data from the tutor for the aforementioned feature analysis. In addition, I converted the interface of the genetics tutor from Flash to HTML5 and also updated the Behavior Recording Diagrams that specify how the tutor functions.

BetterSpeech aims to help people communicate more effectively and improve their public speaking abilities with the added convenience of being able to do so through a website or app. Users upload a video of themselves speaking in order to receive individualized feedback on their speech patterns as well as physical movement/actions. They will also receive a detailed development plan to help them improve. This information comes from a combination of the feedback from other users, professional trainers, and machine analysis.

I worked at the Kumon Learning Center, tutoring children ages 6-14 every day in both math and reading. I also grade their homework and tests, explaining the mistakes that they made to them. Since I have personally also completed the curriculum at the Kumon centers, I was able to offer to the children a more in-depth explanation for how to think through and answer the problems.