CI/CDAndroidIOSMicrosoft Store
Multi-Platform Mobile App Release Pipeline — Android, iOS & Microsoft Store
Multi-stage Azure DevOps CI/CD releasing seven mobile app variants across four stores, with signed builds promoted through platform-specific release tracks on a single version code. Android and iOS live in production; Microsoft Store automation in progress.
80%+ faster releasesIn Progress
// key highlights
[01]
Designed and built a 7-variant, multi-platform release pipeline end-to-end, solo, in Azure DevOps
[02]
Parallelized CI builds across multiple jobs, cutting release time by more than 80%
[03]
Automated signing and artifact promotion across Google Play, Huawei AppGallery, App Store Connect, and Microsoft Store
[04]
Eliminated manual store uploads and manual production promotion for Android and iOS releases
[05]
Mapped Dev/QA/Prod to Google Play's Internal/Closed/Production tracks using single-artifact promotion
// overview
Release engineering for a mobile product shipping 7 app variants, 4 Android configurations, 2 iOS configurations, and 1 Microsoft Store build, from a single monorepo across Google Play, Huawei AppGallery, the Apple App Store, and Microsoft Store.
I designed and built this pipeline entirely on my own: the multi-job CI pipeline building every variant in parallel, the signing and packaging automation for each platform, and the CD pipeline promoting a single tested artifact from initial store submission through to production release.
// problem
The organization's release process for this product was entirely manual. Developers built the Android, iOS, and Microsoft Store apps on their own laptops, then uploaded the resulting build artifacts directly to each platform's store console as a testing release and assigned them to a group of users for QA.
Once QA confirmed a build was ready, that same artifact was manually promoted to production. The entire cycle took more than 4 hours per release, with building the apps on a developer's laptop the single most time-consuming step.
// approach
Started by learning how to build each app locally, seven variants total: 4 Android builds across different configuration profiles, 2 iOS builds across different configuration profiles, and 1 Microsoft Store build, documenting every step, including how each app was signed with its release key.
Studied how to reproduce those manual steps inside an Azure DevOps pipeline: extracting signing keys and uploading them as secure files, installing the right dependencies per platform, and replicating the build and signing process in pipeline form.
Designed a CI pipeline that builds all 7 variants across multiple parallel jobs, publishing every resulting artifact to Azure DevOps Pipeline Artifacts.
Configured service connections between Azure DevOps and each platform (Google Play Console, Huawei AppGallery, App Store Connect, Microsoft Partner Center) so the CD pipeline could push builds directly to each store.
Designed the CD pipeline to publish each build to its platform as a testing release first, with a separate stage to promote that same artifact to production once testing passed.
For Google Play specifically, mapped Dev/QA/Prod to Internal, Closed, and Production tracks, promoting the same signed artifact and version code across tracks without rebuilding.
// architecture
// outcome
End-to-end release time dropped from more than 4 hours to under 45 minutes, more than an 80% reduction, driven largely by building all 7 app variants in parallel across multiple CI jobs instead of one at a time on a developer's laptop. Manual signing, manual uploads, and manual promotion between testing and production were eliminated entirely, replaced by a single artifact built once and promoted unchanged through to release.
Android and iOS are live in production across Google Play, Huawei AppGallery, and the Apple App Store. Microsoft Store release automation is the next phase on the same pipeline architecture.
// references
Internal project, no public references.