* Built a DBT toolkit for rapidly creating uniform voter lists from disparate datasets* Designed, implemented, and measured dozens of RCT experiments* Crafted targeting maps using radio data, voter files, and more

Managed a team of volunteers building dashboards and tools for progressive campaigns and organizations* Worked with a campaign to identify a path to victory and simulate election outcomes* Produced an LLM chat bot to train and engage potential canvassers

* Built tooling for artists to schedule and share social posts to millions of total followers* Created a data warehouse using Glue, S3, PySpark, and Athena* Trained models to cluster art based on a variety of aesthetic dimensions and predict network formation* Built Rails and Django APIs deployed in AWS* Bootstrapped a product that had hundreds of active users and significant MRR

* Led team to build a data pipeline from the ground up* Trained and tuned NLP models to extract structured information from millions of news articles, Tweets, and Reddit posts* Calculated technical indicators on price data and modeled likely price movement

* Designed and implemented pipeline for building and maintaining a knowledge graph and generating free-text insights* Evaluated and implemented DAG schedulers including Airflow and Luigi

* Built containerized streaming summary services to process thousands of records per second and oversaw Kafka, Zookeeper, and other components of a data pipeline* Wrote daily Spark SQL summaries of millions of rows of data * Produced models for clustering users, predicting churn, and identifying changes in service performance.* Worked directly with internal customers to interpret data.

* Designed and deployed anomaly and peak detection algorithms for massive unstructured data sets* Containerized NLP services* Worked directly with customers and subject-matter experts to display linguistic data in an intuitive manner

* Worked with researchers at Argonne to model edge formation in a graph model of grant money in order to predict successful grant applications using a subject-author model to extract features from paper abstracts* Built crawlers to scrape data from millions of sites* Using NLP techniques, built classifiers and browser plugin to detect scam websites using this data.

Supervised a group of approximately ten writers, coordinated with developers, established social-media performance goals and worked towards them

Tested web-based applications, wrote travel-guide content, coordinated email blasts

Proofread and edited textbooks for publication

Planned lessons and taught courses on SAT preparation

Established and maintained a student-run magazine covering music and the arts in the Chapel Hill area