Designed and built the entire QA automation framework from scratch for a SaaS vehicle commerce platform on AWS, using Playwright + TypeScript + Cucumber/BDD. Established a scalable, deterministic testing architecture with strong CI integration, enabling reliable, fast feedback across critical business flows.
Problem: No existing automation framework or quality strategy in place at project start. Testing was manual, inconsistent, and not scalable, with high risk around complex multi-step flows such as checkout, instalment payments, and guarantor processes. Lack of CI integration and performance validation created significant release risk.
Architecture→Designed a Playwright + TypeScript + Cucumber framework from first principles, enforcing separation between step definitions, domain logic, and reusable test utilities for long-term maintainability
→Implemented an API-first validation approach to reduce reliance on brittle UI assertions, using UI tests only for critical user journeys
→Established deterministic test data handling and environment setup to ensure consistent execution across local and CI environments
→Built contract testing capabilities via Smart Pact, enabling automated generation and validation of service contracts to reduce integration risk
→Integrated test execution into GitHub Actions with parallelisation, staged pipelines, and automated quality gates for release decisioning
→Developed k6-based performance testing layer to simulate production-like traffic and validate system behaviour under load
→Structured the framework to support multi-layer testing (API, UI, contract, performance) within a unified ecosystem
Achievements→Achieved 90%+ regression coverage across critical business flows including checkout, instalment payments, guarantor workflows, and admin portal operations
→Created Smart Pact (npm), an open-source library that auto-generates Pact contract tests, reducing test authoring effort by ~60% and standardising service integration validation
→Enabled fast, reliable CI pipelines through parallel Playwright execution and automated Go/No-Go quality gates
→Identified and resolved two critical backend performance bottlenecks using k6 load testing prior to production launch
→Reduced manual testing dependency by introducing a fully automated, scalable test framework from scratch
→Improved overall release confidence by introducing structured, multi-layer validation across the system