AI Photobooths

AI-driven photobooth kiosks deployed at major Indian airport hubs and brand venues. Travellers and event attendees walk up, get photographed, pick a style, and receive an AI-generated portrait themed around iconic local visuals or branded experiences. Several deployments live across multiple venue types.

Role

Solo software, deployed installations

Status

● Deployed at major airport hubs

Duration

2024 – 2026

Form

Touchscreen kiosks with live capture

Venues

Airport terminals and brand events

What it is

AI photobooths are vertical-screen touchscreen kiosks that integrate live photo capture, on-device or cloud AI portrait generation, and themed output. A person walks up, the kiosk captures their image, they select a style or theme, and a stylised AI-composited portrait is returned in seconds. Designed for unattended use in high-traffic public spaces.

AI photobooth deployment at Delhi Indira Gandhi International Airport (Terminal 1, GMR-operated). Player selects a "photostyle" themed around iconic airport visuals; the AI generates a portrait composited into the scene.

The airport installation

The Terminal 1 photobooth offers three photostyle themes: Airport Terminal (the new T1 architecture), Sound of Silence (an iconic sculpture at the airport), and Life's Journey (cultural Indian artwork). A traveller stands in front of the kiosk, taps a style, the camera fires, the AI runs, and within a few seconds they see themselves composited into the selected scene. They can scan a QR code to receive the image.

Style selection screen. Three themed options framed around iconic local visuals.

Live capture. Visitor is photographed against the kiosk's integrated lighting.

Result. Visitor composited into the "Life's Journey" theme with the painted elephant artwork.

The pledge kiosk

A different deployment, same underlying platform. The "Take Your Pledge" kiosk was built around Indian patriotic theming, Independence Day and Republic Day brand activations at GMR airport plazas. Visitors fill in their name, email, and phone, take a photo, and the system returns a stylised personalised pledge card that they can share or print.

"Take Your Pledge" kiosk at a GMR-operated airport plaza. Form capture, photo capture, themed personalised output.

What makes this hard

On the surface "photobooth" sounds trivial. The actual engineering work is in everything around the AI:

Unattended robustness. No staff member is going to reset the kiosk if it crashes at 2 AM in an airport. The application has to recover from network failures, camera disconnects, AI service timeouts, and disk-full conditions automatically.
Image pipeline timing. From "tap capture" to "result shown" has to be under ten seconds. Otherwise the line behind you starts walking away.
Output quality is the entire product. A bad AI portrait is worse than no photo. The system has to handle bad lighting, unusual poses, multiple people in frame, and partial occlusions and still return something the visitor will share.
Privacy. Photos are captured on a public kiosk. They have to be deleted, retention controlled, and access logged.

Stack

Unity / native touchscreen UIImage capture pipelineAI image generation (cloud API)QR-based deliveryKiosk hardware integration

What this project demonstrates

AI-in-production deployment. Image generation hosted somewhere; my job is to make it land in a public installation reliably.
High-traffic public installation discipline. Recovery, retry, queueing, throttling.
Multi-format brand work. Same platform retargeted to different venues, themes, and audiences.