Ellucian

Context

Ellucian builds software for higher education. I worked in research and development on a scheduling product that turned disparate institutional datasets into course schedules for advisors and students.

The work

• Created a production prediction model in Python forecasting student academic performance. It became the data layer behind the schedule-builder used by advisors and students across institutional customers.
• Shipped a scheduling product adopted by the State of Georgia across its higher-education system, generating student schedules from state-provided datasets under strict data-handling requirements.
• Maintained the ML data tables on AWS that supported institutional analytics, including migration and integration across sources.
• Wrote the client-facing web application, on Flask and FastAPI, that put model output in front of non-technical stakeholders as decisions they could act on.

Why it still matters

This was my first time shipping machine learning to a customer who would never look at the model, only at what it told them. That framing carried straight into the AI work at Luxxera, where the output had to be useful to a clinic coordinator before it was interesting to an engineer.