Omid Ameri / PortfolioThe Lady of AVRD
Luna brings voice conversation, a character presence and generative tools into mixed reality on Quest 3. Developed by a six-person Expanded Realities team, the project connects a Unity frontend to a local AI system.
Spatial interaction
I developed the orb menu, mode switching, microphone interaction and visual feedback in Unity. Conversation, image generation, 3D generation and mind-mapping use a shared interaction language.
The four modes share the same orb interaction pattern. Keeping mode selection and recording feedback together gives each tool a familiar entry point.
Application integration
I worked on the real-time WebSocket connection between Quest and the AI backend. Voice input, processing and responses needed to feel like one continuous interaction, with clear feedback while the system was working.
The user-facing application introduces Luna and the controls through onboarding, with a menu that follows the user. Interface states make recording, mode changes and first-time interaction understandable inside the headset.
Local infrastructure
My work also included the Linux / Docker server environment, local LLM deployment and the speech pipeline. The shared backend combines STT, TTS, memory and RAG with separate image and 3D generation workflows.
Speech, memory and generative workflows sit behind one spatial frontend. The architecture keeps the character experience connected while allowing image and 3D tools to extend what the companion can do.
A collaborative six-person project. The development text describes my contributions; the visuals show the shared team result.
All projects
Opening the studio