iOS Case Study: PhotoWash
PhotoWash helps users audit camera rolls and surface sensitive images before they are shared. I led the end-to-end delivery of the app, taking it from concept to App Store launch with a small, focused team.
Brief & Constraints
- Serve the LGBTQI+ community with a private, on-device way to manage NSFW media.
- Ship an accessible SwiftUI interface that feels native on iPhone and iPad.
- Keep all image processing local so no personal data leaves the device.
Approach
- Product design: Prototyped flows in Figma to validate how users scan, review, and bulk-manage flagged photos.
- Core ML pipeline: Trained a custom classifier using a Python-based curation workflow and converted the model for Core ML.
- SwiftUI build: Implemented the experience in Swift and SwiftUI, synchronising with existing Objective-C modules where needed.
- App Store readiness: Automated screenshots, privacy copy, and notarised builds for a streamlined release.
Outcomes
- Accurate detection of medical and adult imagery with on-device inference measured against QA benchmarks.
- Support for smart albums, drag-and-drop, and haptic feedback that aligns with Apple Human Interface Guidelines.
- Positive launch reception, providing a safe workflow for users who need to keep personal images personal.
Tools & Integrations
- Languages: Swift, SwiftUI.
- Frameworks: Core ML, Vision, Combine.
- Automation: Fastlane, GitHub Actions, TestFlight rollout.
