Things I’ve built

A virtual 40k board over a real table. Move the board into place and use terrain layouts to set up a game without measuring every position by hand.

• Reduced my table setup from over an hour to 15 minutes.

Import recipes from social posts and web pages, then cook with step-by-step instructions in your glasses. Talk to the assistant while your hands are busy.

• Built the recipe import pipeline, web app, and smart-glasses cooking experience.

An experiment in turning a real living room into a playable horror level. Exploring room geometry, occlusion, lighting, and enemy movement through the Quest browser.

Drum notation brought into the space around a kit. A browser-based AR prototype that puts visual cues where you play.

• Built and shipped the initial prototype in under an hour.

Flashcards for Meta Display glasses. Ask for a deck out loud, let the app build it, and study without taking your phone out.

An opt-in business-card experience for Meta Display glasses. Recognize other registered users and see the profile information they chose to share.

A focus-session experiment for Meta Display glasses, using IMU signals to explore quiet cues during everyday tasks.

An immersive garden experiment with hand interactions, plants you can pick up, and a journal in the space around you.

A game built for Meta Ray-Ban Display glasses, featured in a Lenslist developer spotlight.

A browser-based sci-fi horror game with a PSX-inspired look, procedural sectors, puzzles, and enemies that respond to sight and sound.

Webcam-based fatigue prototype using facial and eye landmarks to estimate blink behavior and eye-closure dynamics under noisy real-world conditions.

• Developed robust signal processing for blink and closure metrics from commodity camera input.
• Validated prototype behavior across varied lighting and camera quality constraints.

Integrated WebGazer.js to estimate on-screen gaze without specialized hardware.

• Implemented smoothing and region-based aggregation to map gaze to areas of interest.
• Reduced jitter and improved interpretability for lightweight UX research use-cases.

Real-time interview support tool that surfaces AI-generated follow-up prompts during live user research sessions.

• Connected live transcription and retrieval for in-session prompt generation.
• Improved moderator continuity and depth during rapid qualitative interviews.