Photo kiosks that turn a guest's photo into an AI portrait on the spot, running on hardware in the field.
Client
Hardware startup
Year
2024 – 2026
Role
Core developer next to the founder
Stack
React, TypeScript, Node.js, TanStack Query, AI image generation APIs
Live
Private product, no public link
Problem
A kiosk gets one chance: the guest takes a photo, waits, and either gets a good image in seconds or walks away. The machines stand in venues nobody on the team can reach quickly.
Constraints
Real-time image generation that depends on third-party AI services.
Unattended hardware in venues, on venue internet.
Several products and kiosks to run at the same time.
Decisions
A React kiosk UI with camera capture, a touch keyboard and QR handoff, built for a stranger who has never seen it.
A generation pipeline with queues, retries and live progress, so a slow model call never looks like a frozen screen.
A Node.js gateway in front of the AI providers that keeps API keys, routing and traffic in one place.
A fleet admin panel with analytics, remote access and over-the-air updates.
Result
Kiosks running in the field with 40+ AI transformation styles and remote updates across the fleet.
Need something like this?
Tell us what you need
A few sentences are enough to start: what it should do, who uses it, and when you need it. We answer with questions or a price.
Elmo Tech is run by Elmir Beširović, a software engineer who has been shipping to production since 2023. Small team, direct line, no account managers in between.
