Dave
Data Engineer • May, 2023 —Present
Engineer supporting the data science platform. Building and managing tools to train, test, and deploy new models in support of core business services.
More to come as I venture through my first non-research job!
Fortive
Data Engineer • Nov, 2021 —May, 2023
- In order to provide new sources of business intelligence for a marketing funnel, built an ETL pipeline in Prefect and Snowflake which aggregated multiple external and internal data sources to enable the sales team to drive $35M in new sales.
- In support of a legacy customer churn for a large operating company, optimized the various recipes within a Dataiku project to have a 72x improvement in speed and patched a memory leak which required an ever-increasing amount of memory to execute.
- As a core member of the MLOps team, developed Python coding standards, data governance policies, and performed daily management tasks of AWS, Azure, and GitLab tenancy including CI/CD pipelines.
Carnegie Mellon University
Data Engineer • June, 2019 —Oct, 2021
- Developed a bare-metal, ITAR-compliant ETL pipeline utilizing Airflow, Kubernetes, and Python capable of processing a petabyte of custom ROS data from robotic platforms.
- Implemented stereo vision and object-detection algorithms in C++ using Tensorflow and OpenCV.
Noblis
Software Engineer • June, 2017 —June, 2019
- To enable national security analysts to process large amounts of mixed-multimedia, lead the development and research for image and video automatic captioning web application utilizing state- of-the-art ML algorithms.
- In support of a multi-year IARPA research program, refactored legacy Python codebase to facilitate the release and evaluation of HITs on Amazon Mechanical Turk with a 10x performance increase.
- For the Department of Homeland Security’s Kaggle challenge to improve the accuracy of millimeter- wave body scanner threat detection algorithm, developed a convex-hull data preprocessing utility and evaluation harness for a state-of-the-art algorithm which ultimately placed within the top 6%.