Experience-based architecture overview
Mobile and backend workflows across multiple operational sites
A founder-led engineering account of authenticated mobile/API workflows and offline recovery across multiple operating sites.
01
Swift / Kotlin client
02
Authenticated API
03
Server-side policy
Client state cannot grant access
Client/server boundaries for authentication and data flow.
RESOURCES / ENGINEERING STORIES
The problem, the change and the evidence.
An anonymized account of the founder's professional mobile and backend engineering work, with private architecture and identifying details omitted.
A technical account of completed mobile/backend implementation and recovery work, with the original customer system anonymized.
Explore the system, engineering decisions and observed outcome in the sections below.
Project context: An anonymized account of the founder's professional mobile and backend engineering work, with private architecture and identifying details omitted.
Architecture and trust boundaries
A reviewable system path
Conceptual engineering overview — not the customer's actual proprietary architecture.
STEP 1 / 4
Mobile UI and offline state
STEP 2 / 4
Authenticated API boundary
STEP 3 / 4
Server-side authorization
STEP 4 / 4
Authoritative state and audit events
Trust boundary / failure mode
Untrusted client state never grants authorization. Retries require idempotency and server-held freshness.
What evidence supports it?
Representative test cases for discussing scope, architecture and failure modes.
Engineering walkthrough
A short original visual explainer of the technical problem and the control boundary involved.
Problem
- Mobile operations spanning multiple sites have to cope with network interruption, duplicated submissions and inconsistent server state.
- Client UI state cannot be the authoritative record of a sensitive action or its authorization.
Engineering work
- Built and integrated native mobile clients with authenticated REST/JSON services in a multi-site operating environment.
- Used explicit server-side request handling and recovery-oriented workflows so disconnected clients could rejoin normal operation.
- Treated API contracts, data boundaries and operational debugging as part of the delivery, not just the mobile interface.
Engineering validation
- Relevant validation areas include replayed requests, interrupted submissions, out-of-order responses, stale sessions and unauthorized object access.
- The original test records are private. Replay, interrupted writes, stale sessions and cross-object access are the failure cases that define the integration boundary.
Results
- A technical account of completed mobile/backend implementation and recovery work, with the original customer system anonymized.
- Client-specific metrics and internal diagrams remain private.
Related engineering stories
First-party operational work
Repairing a Cloudflare Worker canonical-URL redirect loop
Read case study →
Internal demonstrator
SiteGuard Monitor: service-health and alerting demonstrator
Read case study →
