AZA Finance · QA Engineer · Mar 2020 – Sep 2024
The customer problem
AZA operates payment and FX platforms where a bug isn’t an inconvenience — it’s someone’s money in flight. Incidents originated anywhere: API timeouts, integration failures, deployment misconfigurations, database constraints. The team was strong at reacting; the gap was preventing the second occurrence.
What I investigated
Deep root-cause analysis on every significant incident: production logs, staging environments, release changes, and testing gaps. I mapped each issue to its origin — code logic, test coverage, requirements, design, or environment — which turned “bugs happen” into a pattern library we could act on.
What I built and shipped
- A regression and API validation suite (Cypress, Playwright, k6, Rest Assured, Postman) integrated into GitHub Actions and Jenkins, so every release carried proof it held.
- Comprehensive test documentation — plans, scenarios, acceptance criteria, regression strategies — that outlived individual team members.
- Preventive QA practices that caught reliability risks before deployment rather than after customer reports.
The complexity wasn’t any single test. It was making release validation reliable across interconnected backend services, deciding what to automate versus leave manual, and keeping the suite fast enough that developers actually trusted it.
Outcome: production bug leakage down 30% — a 30% reduction in customer-reported incidents; technical debt reduced by 50% through earlier detection; regression reliability strengthened across backend and API validation.
What I’d tell a customer success team
Every prevented defect is a conversation a CSM never has to have. The cheapest retention program is a release process that stops the bug before the customer’s workflow does.