AI / Education · May 2025 — May 2026
GradeLoop AI
Bringing structure to technical feedback.
Founder & Product Developer
THE WORK
From idea
to practice.
I founded and developed GradeLoop AI, a multi-agent platform for evaluating technical coursework. The work sat alongside my experience supporting data science students and faculty at Boston University.
The problem
Technical coursework brings together code, analytical reasoning and written explanations. GradeLoop focused on bringing those pieces into an AI-assisted evaluation process for educators.
What I built
I engineered a multi-agent grading system and the full-stack SaaS platform around it, with security and auditability included in the design. My work covered both the evaluation system and the product that would make it usable.
Designing for scale
The architecture targeted more than 10,000 grading tasks per day. That figure describes a design target, rather than a measured volume of real student submissions.
The connection to my work
At BU, I evaluated more than 100 technical assignments across 12 courses and worked with over 20 faculty and programme staff. That experience gave me direct exposure to the feedback and assessment workflows this project addressed.