PolicyEngine
PolicyEngine is a non-profit application that models proposed policies and is used by policy makers and academics. See the website.
Explain with AI Feature:
Context: As part of the Policy2Code challenge, PolicyEngine created an “Explain with AI” feature available in each simulation’s result which explains the factors that led to the result. This won the Outstanding Technology Innovation Award.
Solution: I designed the data flow of the feature which works by making a request to an internal API endpoint that returns the computation log of the simulation, passes a prompt to the Anthropic REST API endpoint that includes the log and what output is desired (including format), and displays the response to the user. I also wrote the code for the front-end experience including the connection to both APIs.
Technology: React, Python
Fixed Navigation:
Context: An update to the main application broke the navigation logic in the mobile app. Additional layers were added to the tree that the mobile logic was not equipped to process in order to populate pages when either a navigation button or breadcrumb was clicked.
Solution: Without any context or onboarding, I identified the cause of the problem and implemented a solution. To fix the navigation buttons, I added logic that checks if the next item in the tree has a page associated with it, and if not, it moves to the next item in the tree. For the breadcrumbs, I added logic that checks if the tree item has children, and if so, it presents the children as clickable buttons to the user. I also improved the UX by removing a redundant carousel.
Technology: React