• Used Ruby on Rails and React to integrate Pendo, a third-party product experience software, into the existing Socrata platform.• Developed a suite of configuration options in the Socrata back-end that allow for permission groups and seamless opt-in and opt-out of the software, all linked to a front-end modal placed in the Socrata administration settings.• Developed unit and integration test suites in RSpec, Karma (Enzyme + Chai), and Capybara (Selenium).

• Created a neural network-based image classifier as a proof of concept for image-based fraud detection.• Used PyTorch and Flask for the machine learning models deployed on Paperspace (GPU-as-a-service) as a back-end API. Developed a simple HTML/CSS/JavaScript front-end deployed on Heroku to integrate with the API.• The application involved the use of transfer learning and neural networks to correctly classify different image types.

• Identified cost-saving measures related to scaling or shutting down staging and development servers in periods of inactivity or low usage.• Used Ruby on Rails and Heroku to create and deploy a centralized dashboard web application to manage applications and services.• Centralization and automation efforts resulted in reduced server utilization, helping to save the company $50,000 in hosting costs.• Successfully migrated approximately 12 developers using a combination of documentation, instructions, and direct support.