- Designed and built a real-time, distributed system which alerts operations teams of delivery issues across millions of shipments per day. The system handles ~20k transactions per second to track shipment status- Helped design and implement the acceleration server backing AWS Elastic Inference (https://aws.amazon.com/machine-learning/elastic-inference). Helped launch the product from scratch within 5 months- Implemented a Q&A board using DynamoDB and Node.js, which improvesassociate experience and decreases manager workload and is deployed across all of Amazon’s warehouses- Built an application for surge pay approval for associate shifts, leading to higher fill rates and better visibility into this cost

- Created an on-demand food delivery business for our university with ~10% of the campus using it in the first month

- Using Node.js and MongoDB, built an app which allowed customers to book excess storage space.- Leveraged Elasticsearch to build out low-latency search and filtering capabilities for listings and used Kafka for MongoDB to ES replication

- Worked on the Alexa Skills Store team- Added UI/server functionality in Spring MVC and Backbone.js to change the language in which voice commands appeared, both on Amazon.com and the Alexa app and added monitoring for the feature to measure usefulness- Improved TTFB by 20% by lazily loading images on the Alexa skill store home page- Extended an internal tool to improve efficiency of operations (such as modifying configurations) during high severity incidents

- Selected by Chair of the Mathematics department for tutoring students in mathematics- Tutored students in Java and C programming, Calculus II, and Discrete Mathematics

- Built a desktop email client using the Gmail API- Mitigated searching/retrieval latency by implementing a local cache with SQLite, decreasing latency by 3x

- Tested production code for defects across multiple browsers and operating systems using virtual machines- Integrated an analytics framework (Mixpanel) and built funnels to learn more about how users use the site