AI + XR onboarding
Janu XR
An AI + XR onboarding prototype for first-time headset setup.
First-time headset failures are often physical or procedural: fit, eye calibration, and gestures, but devices rarely explain what a person should change.
An eight-step conversational guide with detailed failure tracking on the three hardest steps (fit, eye calibration, pinch), where three accumulated signals trigger a step-specific intervention. A FastAPI service, a three-screen Jetpack Compose dashboard (Flow, Insights, Replay), and a 39-second Unity demo of KDTree-driven contextual gaze guidance complete the loop.
It connects human-centered AI with observable UX evidence, while staying honest that the Unity gaze and calibration events are simulated rather than live headset telemetry.
Key results
- Eight-step guided setup state machine
- Step-specific failure interventions for fit, calibration, and pinch gestures
- Session persistence and structured evaluation logs
- Android Flow, Insights, and Replay views
- KDTree nearest-anchor selection in the Unity demo
Tech stack
PythonFastAPIGeminiJetpack ComposeUnityC#
Project media
Evidence from the build.
