Jeans

AI-assisted photo editing and sharing service designed for senior users.

2024.12 - 2025.02

Figure 1. Jeans main screen. Primary service interface for AI-assisted photo editing and sharing.

Jeans is an AI-powered photo editing and sharing service designed to make digital photo workflows more accessible for senior users. The project combines voice-based interaction, face-aware image processing, and a mobile-friendly service flow to help older adults edit and share photos with less friction.

Role

PM and AI Lead. Led the AI pipeline as the sole AI engineer, covering dataset preparation, model fine-tuning, API design, deployment, and senior-centered usability validation.

Key Features

  • Voice-driven photo editing using Whisper fine-tuned for Korean dialect and senior speech patterns.
  • Face detection and enhancement with YOLO and FaceNet.
  • Automated captioning and tagging to improve accessibility.
  • FastAPI-based inference APIs deployed on AWS EC2.
  • CI/CD automation with GitHub Actions.
Figure 2. Key features. Overview of voice-driven editing, face-aware enhancement, and photo sharing support.

AI Flow

The service flow connects voice commands, image understanding, and editing actions so that users can complete photo tasks through a simpler interaction model.

Figure 3. Service AI flow. End-to-end flow connecting voice commands, image analysis, and editing actions.

Model Development

Figure 4. Model development. Whisper fine-tuning and dataset preparation for Korean senior speech commands.

Service Architecture

Figure 5. Service architecture. FastAPI-based AI inference services integrated with the mobile photo-editing workflow.

Technical Scope

  • Built speech-command data pipelines with preprocessing for normalization, trimming, and silence removal.
  • Mapped recognized voice commands to image editing operations such as cropping, brightening, and sharpening.
  • Designed REST endpoints for image upload, transformation, and retrieval.
  • Refined the interaction flow through interviews and usability tests with senior users.

Awards

  • Personal Excellence Award, SK Telecom FLY AI Challenger Program.
  • Project Excellence Award, SK Telecom FLY AI Challenger Program.