roadmap
- 01
- 02
- 03
- 04
- 05
- 06
- 07
- 08
Prepare for Playwright interview questions on how its tests find, wait for and verify what a page shows. Covers locators and auto-waiting, fixtures, network interception, tracing and parallel runs.
287 of 287 questions in your plan: 0 known, 0 in progress, 0 due, 0 skipped, 287 new. 0 cut, outside your plan.
QA and SDET loops treat Playwright as the thing you would build and then maintain a suite on, so the questions come from that suite's bad days: why a test goes red only in the pipeline, what should replace a fixed `waitForTimeout`, how to sign in once and reuse the session across the whole run, how to read the trace of the one test that failed overnight. Many of these loops add a live exercise, automating a flow on a demo site and then defending each locator you chose. Candidates who come from Selenium are usually asked what auto-waiting removes from their old habits, and where an explicit wait still belongs. A [Java SDET](/roadmaps/java-sdet) loop adds a language seam. The Java binding ships no test runner of its own, so isolation between tests, parallel execution and retries come from the runner around it, usually JUnit or TestNG, and interviewers check that you know which parts of Playwright travel across languages and which stay with its Node runner. Full-stack developers meet Playwright as the end-to-end layer of their own application, and their questions lean toward the code under test: stubbing the backend with `page.route` so a test does not depend on staging data, and deciding which behavior earns a browser test at all. Playwright is usually one round of those loops, not the whole of it; what the other rounds ask is gathered in the [QA Engineer](/roadmaps/qa-engineer) roadmap and, for the full-stack loop, in [JS Full Stack Developer](/roadmaps/javascript-full-stack).
Let the [Playwright primer](/topics/dev-playwright) give you the map of the tool in one sitting, then spend the hours answering questions rather than reading. Most answers here are judged on one thing: whether you can say what Playwright does between two lines of your test. Junior candidates are expected to get [locators and auto-waiting](/topics/dev-playwright-locators) right before anything else: why `getByRole` holds up where a long CSS path breaks, which actionability checks run before a click, and why `expect(locator).toHaveText()` retries while reading the text once and comparing it does not. That last difference sits behind many of the flaky tests candidates are handed to fix. From middle level up, [fixtures and test configuration](/topics/dev-playwright-fixtures-config) start to separate people: a custom fixture with its own setup and teardown, a worker-scoped one for state too expensive to rebuild per test, a saved signed-in state that replaces the login step everywhere. At senior level the weight moves to [run orchestration](/topics/dev-playwright-execution): sharding across CI machines, what retries hide, when `fullyParallel` breaks tests that share data, and which traces are worth keeping from a red run. Close to the date, raise importance to must know but leave level on its full range, since one flaky test can carry a junior question and a senior one, and what changes between them is how far down the answer has to go.
QA and SDET loops treat Playwright as the thing you would build and then maintain a suite on, so the questions come from that suite's bad days: why a test goes red only in the pipeline, what should replace a fixed waitForTimeout, how to sign in once and reuse the session across the whole run, how to read the trace of the one test that failed overnight. Many of these loops add a live exercise, automating a flow on a demo site and then defending each locator you chose. Candidates who come from Selenium are usually asked what auto-waiting removes from their old habits, and where an explicit wait still belongs.
A Java SDET loop adds a language seam. The Java binding ships no test runner of its own, so isolation between tests, parallel execution and retries come from the runner around it, usually JUnit or TestNG, and interviewers check that you know which parts of Playwright travel across languages and which stay with its Node runner. Full-stack developers meet Playwright as the end-to-end layer of their own application, and their questions lean toward the code under test: stubbing the backend with page.route so a test does not depend on staging data, and deciding which behavior earns a browser test at all. Playwright is usually one round of those loops, not the whole of it; what the other rounds ask is gathered in the QA Engineer roadmap and, for the full-stack loop, in JS Full Stack Developer.
Let the Playwright primer give you the map of the tool in one sitting, then spend the hours answering questions rather than reading. Most answers here are judged on one thing: whether you can say what Playwright does between two lines of your test.
Junior candidates are expected to get locators and auto-waiting right before anything else: why getByRole holds up where a long CSS path breaks, which actionability checks run before a click, and why expect(locator).toHaveText() retries while reading the text once and comparing it does not. That last difference sits behind many of the flaky tests candidates are handed to fix. From middle level up, fixtures and test configuration start to separate people: a custom fixture with its own setup and teardown, a worker-scoped one for state too expensive to rebuild per test, a saved signed-in state that replaces the login step everywhere. At senior level the weight moves to run orchestration: sharding across CI machines, what retries hide, when fullyParallel breaks tests that share data, and which traces are worth keeping from a red run.
Close to the date, raise importance to must know but leave level on its full range, since one flaky test can carry a junior question and a senior one, and what changes between them is how far down the answer has to go.