Case Studies

Selected work

Four engagements from roughly a year of test automation work on a multi-tenant commerce operations platform: a React micro-frontend application, a backend-for-frontend, and several GraphQL subgraphs behind a federated router.

Client, product, service and people names have been removed, and figures are counted from version control rather than estimated. We will go deeper on any of these on a call, under NDA where that helps.

Building a UI suite that survives its own frontend

end-to-end test cases: 85
page-object classes: 48

85 end-to-end test cases and 48 page-object classes on a multi-tenant admin platform, built to survive two frontend migrations.

Playwright
TypeScript
end-to-end testing
page object model

Read the case study

API contract and authorisation testing on a federated graph

API test cases: 62
call sites for the persistence read-back module: 44

62 API test cases, a persistence read-back module called from 44 sites, and 22 silent skips turned into hard failures.

GraphQL
federated graph
API testing
authorisation testing

Read the case study

Taking over a red pipeline

quarantined tests restored in 20 days: 12
leaked records per run: 7 → 0

12 quarantined tests restored in 20 days, leaked test data cut from 7 records per run to zero, and a suite that stopped killing other teams' pipelines.

CI
flaky tests
test data management
pipeline stability

Read the case study

An AI-augmented QA workflow

lines in the custom agent skill: 467
accessibility snapshots over seven months: 796

A custom agent skill, live locator verification through MCP, and an enforcement layer that blocks a commit until the self-audit is clean.

AI tooling
Claude Code
MCP
Playwright
developer workflow

Read the case study