Machine Learning Based Intelligent Test Selection for Faster CI/CD Pipelines

CI pipelines become slow as regression suites grow. In many teams, every commit triggers full test execution even when only a few components changed. In this project, I built a practical prototype that predicts impacted Playwright tests using machine learning. The Problem When all tests run on every commit: feedback is delayed compute cost increases developer productivity drops For large systems, this creates a release bottleneck. The Idea Use historical data from CI: changed files in commit tests that were impacted (failed, flaky, or behaviorally affected) Train a model that maps file-change patterns to impacted test files. Then in CI: detect changed files predict relevant tests run only selected tests first keep full-suite fallback/nightly run for safety Example Commit touches: src/services/inventory.js Model predicts: tests/playwright/tests/inventory.spec.js tests/playwright/tests/order.spec.js This gives much faster feedback compared to running all tests. Tech Stack Playwright for