Rachel Ratner

OlmoEarth: Built a full-stack application enabling researchers to label geospatial data and train machine learning models. Developed the TypeScript/React front end, Flask/Python REST APIs, and database schemas. Trained and fine-tuned ML models and evaluated results. Collected requirements from ML researchers and identified opportunities for new tooling to simplify and scale their workflows.

MolmoSpaces: Built the demo website for MolmoSpaces, which renders MuJoCo physics simulations in the browser. Integrated MuJoCo with Three.js to create navigable 3D environments for virtual robots.

Skylight: Was the primary UI developer on the Skylight project following its transition from Vulcan to AI2. Built UI tools for annotating and visualizing ML training data.

Architected and developed end-to-end applications for the Skylight team, a platform for tracking illegal fishing activity. Built a REST-based OpenAPI and a GraphQL API from the ground up. Developed a real-time React web app for vessel tracking, a back-end notification service for alerting users to illegal fishing events in their area, and managed metrics via Prometheus and Google Analytics. Collaborated closely with conservation stakeholders and cross-functional teams. Note: The Skylight project transitioned from Vulcan to AI2 in 2021.

Developed custom landing pages for Fortune 500 companies, and built a reusable “Landing Page Builder” in React that streamlined the creation and scalability of these pages. Collaborated extensively with QA and designers within the Amazon Marketing Group.

Worked with UX and design teams to build pixel perfect responsive custom web pages for marketing clients.

Worked with a variety of agencies and clients to create custom built, responsive online solutions including a website for Pokémon (via Plexi-Pixel) and a Facebook App for Chateau St. Michelle (Cole & Webber).

Managed KEXP’s marketing team, overseeing website updates, managing internships, and hosting on-air segments.