o Own product usage data end to end as its primary owner: the nine core pipelines that are our source of truth for product usage, 300+ dbt models, and 700+ tests, feeding the executive dashboard the CEO reads every morning, the reporting Customer Success and Sales run on, and the usage numbers customers see in the product

o Define governed metrics in Transform so Customer Success, Sales, Product, and engineering read the same definitions in every tool

o Took on ingestion over the past year: consuming the product streams platform engineers publish to Kafka into Snowflake on Airflow, flattening nested JSON in dbt, and validating new session-level streams against known platform activity, as part of the team that cut data latency from about an hour to under five minutes

o Built shared logic giving those streams a common plan ID and timestamp structure, including session-level plan attribution, and moved them from test into our enterprise Snowflake environment

o Built a human-in-the-loop Claude agent that checks the source topics in our data lake weekly, flags in Slack where a model is reading an older topic version, and opens a pull request with the fix; its first run caught 40+ models reading outdated source topics

o Led both migrations after the A Cloud Guru acquisition: moved about 40 reports from Looker to Tableau and Snowsight with the logic in dbt, and moved about 20 product data streams across roughly 1,500 customers into our enterprise Snowflake environment, owning the parity audits that gated each cutover

o Led the migration of 35 to 40 Customer Success reports off a retiring Snowflake platform: ran discovery with CS leadership, chose Streamlit with no new tool spend, built the template, and delegated the rest to two teammates

o Manage Snowflake column masking policies and data shares in Terraform

o Mentored three new hires: one data engineer and two analytics engineers

o Led organization-wide adoption of dbt after training at a dbt conference, setting the testing, documentation, and modular design standards the analytics team followed

o Designed and implemented the analytics layer of the enterprise data warehouse in dbt: 50+ tested and documented models serving all four Universal Service Fund program areas under FCC oversight

o Owned end-to-end predictive analytics pipelines in R, from fraud detection to propensity to sign up for a service, with Great Expectations data quality checks and performance monitoring

o Built dashboards and reporting in Tableau and R Shiny with program stakeholders, designing the underlying models around the decisions they needed to make

o Integrated legacy source systems into tested star schema data marts so stakeholders could self-serve while data governance held

o Mentored three new analysts on dbt, SQL optimization, and dimensional modeling through code review

o Designed and taught graduate-level courses on data science for public policy students, covering SQL, Python, R, data warehousing, version control, machine learning, and text analysis

o Mentored 60+ students through code reviews and office hours on analytics best practices, reproducible workflows, and technical communication

o Guided students in collecting real-world datasets, designing statistical models, and conducting analyses to inform policy proposals

o Oversaw development of machine learning projects evaluating solutions for public sector challenges, including NYC transit system optimization

o Built data pipelines using Python, Airflow, and PostgreSQL to enable ongoing analysis of publication and citation data, evaluating the university’s academic performance relative to peer institutions

o Built a multi-decade PostgreSQL data warehouse integrating data from sources across the university, primarily the College of Behavioral Sciences, to analyze undergraduate retention for college leadership

o Assembled datasets measuring individual faculty performance across teaching, publication, and service

o Served as a go-to data resource for the Provost, taking on ad hoc data projects for college and university leadership

o Managed data collection for a precinct-level primary election research project spanning 14 states, organizing the data in a warehouse for affiliated scholars

o Collaborated with faculty on dataset collection, analysis, and visualization supporting research publications using R, STATA, and ArcGIS

o Designed course materials, planned lessons, provided constructive feedback to students, and led sections in a range of courses, such as American Government and Politics, Introduction to GIS for Social Science Research, and Advanced Statistical Methods for Social Science

o Developed a remote lab process to increase access and efficiency for students and faculty using ArcGIS

o Created dummy and real datasets to demonstrate complex models and mapping processes in R, STATA, and ArcGIS