All projects

CASE_STUDY // MARKO

marko

marko builds AR maps of indoor spaces like malls, campuses, and hospitals from floor plans and one iPhone LiDAR scan. Visitors say where they want to go, marko localizes them from the camera in about 0.4 s against a map of 7,753 posed photos, and guides them with an on-ground AR path, a Find My-style arrow, and spoken turn-by-turn directions across floors. It took the BigRedTrack prize at BigRed//Hacks 2026.

I built the on-ground AR path in Swift and RealityKit — a GroundPath module that turns routes into floor-hugging geometry, with a destination ring, a green fade-in, and unit tests. On the venue side I built the React analytics dashboard (KPIs, traffic charts, floor map, and 3D scan view), set dark mode as the default with a toggle on a new Settings page, and made the ElevenLabs voice agent opt-in behind an explicit mic button.

The wider stack runs SuperPoint, LightGlue, and NetVLAD localization on rented RTX 4090s, Dijkstra routing over a 0.5 m grid, and a SwiftUI + ARKit app, built with William Sharpe, Neekon Tabatabai, and Mateo del Rio Lanse.

DEMO