← Index

SupplAI

End-to-end regression and security spot-checks on a logistics planning and live fleet-tracking platform.

Client
Codesage
Role
Senior QA Engineer — regression and edge-case coverage
Period
Feb — Mar 2026
modules covered
15modules covered
formal test reports delivered
5formal test reports delivered
live telematics integration
GPSlive telematics integration

Overview

Logistics & fleet management

SupplAI plans journeys, allocates trucks and drivers, and tracks them live over a GPS integration while a control room watches for stops, breakdowns and alerts. It was my first engagement in this role, and the domain punishes shallow testing: a journey has a lifecycle, and most real defects live in the transitions rather than in any single screen.

Journey lifecycleCreatedmanual / bulkConfirmedvalidatedPlannedtruck + driverLiveGPS trackedCompletedETA reconciledArchivedreportedControl room — while liveStop managementBreakdown handlingAlerts & logsETA & time logicA journey screen can pass in isolation and still be wrong — the defects live where one state becomes the next.Free-text fields feeding control-room dashboards were probed for XSS alongside the functional suite.
The journey lifecycle — most defects live in the transitions, not the screens.

01

Coverage

  • Master data administration, journey creation and bulk journey upload
  • Journey confirmation, planning and fleet allocation under division-based rules
  • Live operations tracking, stop management, breakdown and alert handling
  • Control room logs, ETA and time logic, journey completion, archive and reporting

02

What I automated

  • Playwright specs across customers, drivers, fleet and a full end-to-end regression path
  • Filter edge-case coverage, where combinations rather than single filters produced the failures
  • Cross-verification runs reconciling the specification against actual behaviour
  • Automated Excel report generation so results landed in the format the client already used

03

Security spot-checks

Fleet platforms accept a lot of free text — customer names, waypoint labels, log notes — and all of it is rendered back into dashboards. I ran cross-site-scripting probes and whitespace and boundary tests against those inputs alongside the functional suite, on the principle that an input field nobody validated is a defect waiting for a worse discoverer.

Stack

Automation

  • Playwright
  • JavaScript
  • Excel report generators

Domain

  • Journey lifecycle
  • Fleet allocation
  • GPS telematics
  • Control room operations

Practice

  • E2E regression
  • Filter edge cases
  • XSS probes
  • Cross-verification