Data Scientist
2018 - 2019One Concern- Researched and implemented a novel machine learning algorithm for flood inundation.
- Developed a pipeline for land-use classification from satellite imagery.
- Built a ground-truth dataset of historical flood events using satellite imagery.
Technologies: Docker, GDAL, TensorFlowData Scientist
2016 - 2018Retail Solutions- Developed a sales forecasting algorithm incorporating unstructured promotional and sporting event data, as a result, a large customer renewed their contract. The forecasting was done in R, with Python for the ML components, and SQL for ETL.
- Carried out a performance audit of a critical R machine learning pipeline which reduced server usage by 60%, enabling the employer to meet SLAs that they were previously failing.
- Performed ad-hoc investigations and presented results to customers, querying multi-petabyte Vertica SQL and Spark clusters for relevant data.
- Built ETL pipelines for messy data, using Python and SQL.
- Produced interactive visualizations to help clients understand the parameters of their advertising campaigns, with d3.js.
Technologies: Microsoft SQL Server, XGBoost, Vertica, RSoftware Engineer
2015 - 2016MetOcean Solutions- Led a rewrite of the flagship product: a web application to visualize oceanographic forecasts. The new application is currently being used by all customers.
- Designed, built, and deployed a production REST API to interpolate raw weather data.
Technologies: NumPy, Flask, React, D3.js