Fabrio — AI Grading for Engineering CAD at Scale
Rebuilt the Fabrio platform from the ground up to revolutionise grading for engineering CAD work — saving professors thousands of hours at universities like Imperial College London and UCL.
Fabrio is a platform that revolutionised grading for engineering CAD work. Professors at engineering schools were marking thousands of individual Fusion files by hand every term; Fabrio turned that into instant, automated assessment with live progress dashboards, saving them thousands of hours a year.
I joined as the first engineering hire. The company raised venture money, grew to a 6-figure ARR, built a team of six, and secured customers like Imperial College London and University College London — with the platform ultimately reaching 40+ institutions worldwide, from Surrey and Birmingham in the UK to universities across the US, Italy, and Hong Kong, plus programmes like F1 in Schools and Co-op Academies Trust.
I led the ground-up rebuild of the platform on AWS serverless (Cognito, AppSync, Lambda, DynamoDB, S3) with a TypeScript and Next.js frontend over GraphQL. SSR caching took p95 page load from 3.1s to 680ms; the grading engine was rebuilt as a queue-backed worker pool that processed thousands of submissions during deadline peaks without the instructor dashboard ever stalling.
Beyond performance, we made the platform collaborative — live CAD sessions, shared assignments, tight feedback loops — and trustworthy to institutional buyers, with auth, logging, and access controls that held up under university IT review. We also established a pioneering partnership with Autodesk, validating the approach to transforming CAD education.
Saved professors thousands of hours across Imperial College London, UCL, and 40+ institutions; 6-figure ARR with a team of six.