We don't give interns busy work. From week one, you're shipping code, building AR experiences, and solving problems that matter on live client projects.
Choose a discipline that matches your background and ambitions. All tracks work on real deliverables — no sandbox projects.
Build augmented reality experiences using Unity, WebXR, and HoloLens. Work on spatial anchors, image tracking, and interactive 3D scenes for enterprise clients and cultural institutions.
Work on building information modelling pipelines — ingesting IFC/Revit files, building REST APIs, and creating digital twin dashboards that city planners and architects use daily.
Train and deploy ML models for object detection, work with 3D point cloud data, and build real-time inspection dashboards. Hardware + software — from NVIDIA Jetson to browser-based visualisation.
A structured but hands-on program. You'll be mentored closely, but expected to contribute independently from the start.
Week one is deep orientation — codebase walkthrough, design system, client context, and a small starter task scoped to your track. You'll have a technical mentor assigned from day one.
Weeks 2–8: you're on a live project. Real deliverables, real deadlines, real clients. You'll attend project standups, code reviews, and client check-ins. No sandboxed "learning projects".
Weeks 9–10: one focused sprint on something you pitched. A new feature, a tool, a prototype. This is your space to pursue something interesting — and many intern projects have shipped to production.
Final week: demo to the full team. Strong interns are considered for ongoing part-time roles. We don't promise conversion but we've hired from every cohort so far.
I spent my first two weeks refactoring the BIM processing pipeline for MetaBIM. Real code, production codebase. I learned more in those two weeks than in a semester of university projects.
The AR/XR track was intense in the best way. I built an image-tracking prototype using mind-ar.js in week two, and by week six it was in a client demo. That kind of speed is rare anywhere.
I came in as a CV researcher, left with end-to-end production deployment experience — model training, ONNX export, Jetson edge inference, React dashboard. That stack is on every job listing I see now.
A selection of features and tools built by interns that made it to production.
A real-time visualisation tool that highlights structural changes between two versions of an IFC file — now part of the MetaBIM platform core.
View Project →Positional audio system for the Big Plans AR experience — sound zones that activate on proximity to exhibit panels using the Web Audio API.
View Project →A YOLO post-processor that classifies pipeline defects into severity tiers (low/medium/critical) and aggregates scores per pipeline segment.
View Project →Next cohort starts June 2025. Applications close May 9. Three spots across the three tracks.