Rahul Acharya Email & Phone Number

Mountain View, Ca, Us

Backend infra at YouTube

Mountain View, Ca, Us

YouTube Identity Infrastructure

Mclean, Va, Us

- Utilized Swift and Combine to modernize the autopay feature in Capital One's flagship customer iOS app; the new feature is a reactive UI component that enables credit card holders to make or cancel an auto payment instantly; scheduled for deployment in the next app release (fall 2021)- Used Capital One's custom reactive store architecture for deployment; wrote code in Swift to embed business logic into UI component; leveraged Combine to perform network requests to internal APIs and store the responses in custom Swift objects- Partnered with the Design team to optimize customer experience

Seattle, Wa, Us

- Implemented automated, reactive load-balancing tool in Java that locates storage hotspots that form across the distributed system and performs algorithmically generated migrations to rebalance the fleet load; reduced tail latencies faced by Elastic File System customers as well as operational burden by automating the previously manual load-balancing process- Stored configurable algorithm parameters using DynamoDB; wrote selection and migration logic under strict time complexity constraints- Performed exhaustive unit-testing using Mockito; wrote integration tests, deployed tool, and created dashboards to visualize algorithm-generated metrics

New Brunswick, Nj, Us

- Developed an end-to-end micro service that utilizes C5.0 boosted trees to predict the general ledger/commodity code of a given purchase order; used R to implement models and wrapped them in Flask to allow user interaction; wrote integration tests using Tavern- Co-created a RShiny application that visualizes clusters created by an unsupervised machine learning algorithm and delivers insights

New York, New York, Us

- Leveraged machine learning to identify key drivers of patient drug adherence to inform Pfizer’s commercial strategy- Used feature engineering on pharmacy data to intuitively derive relevant patient metrics without creating target leaks- Developed and tuned predictive models in Dataiku to determine key drivers of adherence

Boston, Us

Course Assistant for CS 2800 Logic and Computation (Summer 1 2018)- Maintained office hours twice a week for 34 students- Answered student queries on Piazza (online forum)- Graded homework assignments and exams