Back to work
E-Commerce Unicorn

Vestiaire Collective

I engineered a high-performance Photo Library engine at Vestiaire Collective using SwiftUI and Combine — cutting upload times by 20%, lifting engagement by 15%, and reducing bug reports by 30% through a scalable modular MVVM-C architecture.

Senior iOS Developer·Sep 2020 — Mar 2026·Company
SwiftUICombineMVVM-CFastlaneUIKit
20% faster
Uploads
15% higher
Engagement
30% fewer
Bug reports
50% faster
Deployment

Vestiaire Collective is one of the fashion industry’s leading resale platforms. The Photo Library is where sellers capture, process and publish the images that drive their listings, so performance there has an outsized impact on the marketplace.

I built the Photo Library engine on SwiftUI and Combine, modularizing the pipeline so image capture, processing and display each lived in a clean, testable unit. The result was measurably faster uploads and higher engagement, precisely because the architecture isolated the bottlenecks.

Beyond the feature, I scaled the app’s modular MVVM-C architecture and overhauled CI/CD with Fastlane — cutting deployment time by half and materially reducing escaped bugs.

The challenge

  • Photo uploads were slow and the library UX saw low engagement — a critical flow for a fashion resale platform where images drive listings.
  • The codebase had grown hard to navigate; coordination between many feature teams was slowing delivery and letting regressions through.
  • Deployment cycles were long, keeping the team from shipping improvements quickly.

My approach

  • Architected a modular MVVM-C Photo Library engine in SwiftUI and Combine, isolating image ingestion, processing and display into testable modules.
  • Scaled a modular MVVM-C architecture across the app, tightening boundaries so teams could work in parallel with fewer conflicts.
  • Overhauled CI/CD with Fastlane, automating builds, tests and releases to shorten the path to shipping.
What I learned
  • Isolating pipeline stages into modules both speeds the feature up and makes it testable in isolation.
  • CI/CD investment pays off directly in stability — faster releases with fewer regressions.

Related work