Question 1

Explain the automation framework used in your project.

Accepted Answer

In my last project we used a hybrid framework with Selenium and Java, TestNG as the runner, Maven for builds, and the Page Object Model for UI layers. Test data came from JSON and Excel, reports from Extent, and everything ran in Jenkins on Selenium Grid.

Question 2

How do you decide which test cases to automate?

Accepted Answer

I automate stable, repetitive, high-value cases first: smoke and regression of critical flows, data-driven scenarios, and cross-browser checks. I avoid automating one-time tests, frequently changing UI, and things needing human judgement like usability.

Question 3

What is your day-to-day work as an automation engineer?

Accepted Answer

I start by analysing the nightly run failures and raising bugs or fixing scripts. Then I attend stand-up, work on automating sprint stories, review teammates' pull requests, and help with manual testing of new features when needed.

Question 4

How do you handle dynamic elements whose IDs change?

Accepted Answer

I avoid generated IDs and use stable attributes like data-testid, name, or visible text, with relative XPath or CSS using contains or starts-with. Where possible I ask developers to add test IDs.

Question 5

How do you handle iframes and multiple windows?

Accepted Answer

For iframes I switch into the frame with switchTo().frame and switch back with defaultContent. For new windows or tabs I store the parent handle, loop through getWindowHandles, switch to the new one, and return when done.

Question 6

How do you automate file upload?

Accepted Answer

If the page has an input of type file, I just sendKeys the absolute file path, which works on Grid too with a LocalFileDetector. I avoid Robot class or AutoIt because they break in headless and remote runs.

Question 7

What wait strategy do you follow in your framework?

Accepted Answer

We use explicit waits with WebDriverWait wrapped in reusable helper methods, and we keep implicit wait at zero to avoid mixing them. Thread.sleep is not allowed in code review.

Question 8

How do you handle flaky tests in your project?

Accepted Answer

I first find the root cause rather than just adding retries: usually missing waits, test data collisions, shared state, or environment instability. Flaky tests are quarantined from the release gate until fixed, and we use limited retry only as a short-term safety net.

Question 9

Your tests pass locally but fail in CI. How do you debug it?

Accepted Answer

I compare the environments: headless mode, screen resolution, browser and driver versions, timing on slower agents, test data, and environment URLs. Screenshots, logs, and videos from the CI run usually show the cause quickly.

Question 10

How did you implement parallel execution?

Accepted Answer

We run tests in parallel with TestNG at the methods or classes level, and each thread gets its own WebDriver stored in a ThreadLocal. Tests run on Selenium Grid in Docker, which cut our regression from about 4 hours to 40 minutes.

Question 11

How do you manage test data in automation?

Accepted Answer

We keep static data in JSON or properties per environment and create dynamic data through APIs or database setup before tests, so each test owns its data. Data is cleaned up after the run where possible.

Question 12

What reporting do you use and what does it include?

Accepted Answer

We used Extent Reports, later Allure, attached to TestNG listeners. Each test shows steps, environment, browser, and on failure a screenshot and logs, and the report link is posted to the team channel after each Jenkins run.

Question 13

How are your tests integrated into CI/CD?

Accepted Answer

Smoke tests run on every deployment to QA via a Jenkins pipeline trigger, and the full regression runs nightly. Parameters like environment, browser, and suite are passed to Maven, and failures notify the team on Slack or email.

Question 14

How do you automate flows with OTP or CAPTCHA?

Accepted Answer

CAPTCHA should not be automated; we ask developers to disable it or use a test key in QA environments. For OTP we fetch it from a test API, the database, or a test email/SMS service like Mailosaur, or use a fixed OTP in lower environments.

Question 15

How do you reduce maintenance effort when the UI changes often?

Accepted Answer

Locators live only in page objects, so a UI change means updating one place. We use stable data-testid attributes, reusable components for common widgets, and move business validations to the API layer.

Question 16

How do you estimate automation effort for a feature?

Accepted Answer

I break the feature into scenarios, classify them as simple, medium, or complex based on steps, data, and integrations, and apply our historical averages, for example 2, 4, and 8 hours. Then I add time for new page objects, data setup, review, and stabilisation.

Question 17

How do you measure the value of your automation?

Accepted Answer

I track regression execution time saved, automation coverage of critical flows, pass rate stability, and defects caught by automation before release. In my project, automating regression reduced the cycle from 3 days manual to a few hours.

Question 18

If you had to build a framework from scratch, how would you start?

Accepted Answer

I first understand the application, tech stack, team skills, and CI setup, then choose tools to match. I build a small skeleton with driver management, config, POM, reporting, and CI integration, prove it with a few smoke tests, and then scale.

Question 19

What was the biggest challenge you faced in automation and how did you solve it?

Accepted Answer

Our nightly regression had around 15 percent random failures, so nobody trusted it. I categorised failures over two weeks, found most were shared test accounts and missing waits, introduced per-test data via API and a wait helper, and brought flakiness below 2 percent.

Question 20

When do you use JavascriptExecutor, and how do you handle Shadow DOM?

Accepted Answer

I use JavascriptExecutor only as a fallback, for scrolling or when an element is genuinely intercepted, because JS clicks bypass real user behaviour. For Shadow DOM, Selenium 4 supports getShadowRoot to search inside the shadow tree with CSS.

Question 21

Walk me through the folder structure of your automation project.

Accepted Answer

We separate concerns: pages or components hold locators and actions, tests hold only scenarios and assertions, and utilities, config and test data live in their own folders. Environment config is externalised so the same code runs against QA, staging or prod-like environments.

Question 22

Is your framework data-driven, keyword-driven or hybrid?

Accepted Answer

Most real projects are hybrid: a Page Object Model core, data-driven tests fed from JSON, CSV or Excel, and sometimes a BDD layer on top for business-readable scenarios. Pure keyword-driven frameworks are rarer today because they are harder to debug than plain code.

Question 23

Why did you use the Page Object Model in your project?

Accepted Answer

POM keeps locators and page actions in one class per page or component, so when the UI changes we fix one place instead of dozens of tests. Tests read like user flows and contain no raw locators.

Question 24

How do you validate data in a dynamic web table?

Accepted Answer

I locate the row by a unique business value, such as an order ID, and then read the cell in the required column relative to that row, instead of using fixed row indexes. For paginated tables I either use filters/search or loop through pages until the row is found.

Question 25

How do you automate and verify a file download?

Accepted Answer

In Playwright I wait for the download event, save the file and assert its name and content. In Selenium I set the browser's download directory via options and poll that folder until the file appears and is complete.

Question 26

How do you handle JavaScript alerts and browser popups?

Accepted Answer

In Selenium I switch to the alert with driver.switchTo().alert() and accept, dismiss or read its text. In Playwright dialogs are auto-dismissed by default, so I register a page.on('dialog') handler before the action that triggers it.

Question 27

In a real project, what caused StaleElementReferenceException for you and how did you fix it?

Accepted Answer

It means the element I stored was removed or re-rendered in the DOM. I fix it by re-locating the element right before using it, waiting for the re-render to finish, and avoiding caching WebElements across page updates.

Question 28

How did you cut login time across your suite without losing coverage of the login flow?

Accepted Answer

I log in once via API or a setup step, save the session cookies or token, and inject them into each test's browser context. Only the dedicated login tests go through the UI login form.

Question 29

A regression test failed overnight. What steps do you take in the morning?

Accepted Answer

I check the report, screenshot, logs and trace to see whether it is a product bug, a test issue or an environment problem. I re-run it once to check for flakiness, reproduce manually if needed, then either raise a defect with evidence or fix the script.

Question 30

How do you tell whether a failure is a real bug or a script issue?

Accepted Answer

I reproduce the step manually on the same environment and data. If it works manually, I look at locators, timing and data in the script; if it fails manually, it is a product bug and I log it with steps, evidence and environment details.

Question 31

Do you use retries for failed tests? What are the risks?

Accepted Answer

Yes, we allow one or two retries in CI to absorb rare infrastructure blips, but every test that passes only on retry is reported as flaky and tracked. Retries must never hide a genuine intermittent product bug.

Question 32

How do you do cross-browser testing in your project?

Accepted Answer

We pass the browser as a parameter from config or the CI job, and a driver factory creates the right driver. Smoke runs on all supported browsers, while full regression runs mainly on Chrome with a periodic run on others, often via Selenium Grid or a cloud like BrowserStack.

Question 33

How do you combine API and UI automation in one test?

Accepted Answer

I use APIs for setup and teardown, such as creating a user or order, and use the UI only for the behaviour under test. This makes tests faster and less brittle, and I can verify the UI result back through an API call.

Question 34

How do you run the same suite against different environments?

Accepted Answer

Environment-specific values like base URL, users and DB strings live in config files or environment variables, and we pick the environment with a command-line parameter or CI variable. The test code itself never changes between environments.

Question 35

How do you implement logging in your framework?

Accepted Answer

We use a logging library like Log4j2 or SLF4J in Java, or the reporter's step API in Playwright, to log key actions, data and API responses. Logs are attached to the report so a failure can be understood without re-running.

Question 36

How do you capture screenshots on failure?

Accepted Answer

In TestNG I implement an ITestListener and take a screenshot in onTestFailure, then attach it to the Extent or Allure report. In Playwright I set screenshot: 'only-on-failure' and trace: 'on-first-retry' in the config.

Question 37

You have only 2 hours before release. How do you prioritise regression?

Accepted Answer

I run the automated smoke suite first, then risk-based selection: areas touched by this release, critical revenue flows like login and checkout, and features with recent defects. I tell stakeholders clearly what was covered and what was not.

Question 38

How do you organise smoke, sanity and regression suites?

Accepted Answer

We tag tests rather than duplicating them: @smoke for build verification, @regression for the full set, and feature tags for targeted runs. TestNG groups or XML suites, Cucumber tags, or Playwright grep select them at runtime.

Question 39

A developer says the bug you raised is not a bug. What do you do?

Accepted Answer

I go back to the requirement or acceptance criteria and show the evidence: steps, expected versus actual, screenshots or logs. If the requirement is unclear, I involve the product owner to decide, and we update the story so it is clear next time.

Question 40

Your manager asks you to automate 200 test cases in one sprint. How do you respond?

Accepted Answer

I give a data-based estimate from past velocity, then propose a prioritised plan: automate the highest-risk and most-repeated cases first, and schedule the rest. I make the trade-off visible rather than silently cutting quality.

Question 41

How do you automate when requirements are incomplete or changing?

Accepted Answer

I clarify acceptance criteria in refinement, start with stable parts like APIs, and keep UI automation for after the design settles. I build tests from small reusable steps so changes are cheap.

Question 42

How do you validate network calls or mock backend responses in UI tests?

Accepted Answer

In Playwright I use page.route to intercept and mock responses, and page.waitForResponse to assert real calls. In Selenium 4 I can use the DevTools (CDP) API on Chromium or a proxy, but it is more limited.

Question 43

Why must automated tests be independent, and how do you ensure it?

Accepted Answer

Independent tests can run in any order and in parallel, and one failure doesn't cascade into others. Each test creates its own data, uses a fresh browser session and cleans up after itself.

Question 44

How did you make WebDriver thread-safe for parallel runs?

Accepted Answer

I store the driver in a ThreadLocal inside a DriverManager, so each TestNG thread gets its own instance. Page objects receive the driver from the manager, and I call remove() in teardown to avoid leaks.

Question 45

Have you done visual regression testing? How?

Accepted Answer

We used screenshot comparison, either Playwright's toHaveScreenshot or a tool like Applitools or Percy, on key pages. Baselines are stored per browser and OS, and dynamic regions like dates are masked to avoid false failures.

Question 46

How would you add accessibility checks to your automation?

Accepted Answer

I integrate axe-core, for example @axe-core/playwright, to scan key pages and fail on serious or critical violations. Automated scans catch only part of WCAG issues, so manual keyboard and screen-reader checks are still required.

Question 47

How would you plan a migration from Selenium to Playwright?

Accepted Answer

I start with a proof of concept on a few critical and flaky flows to measure speed and stability gains. Then I migrate incrementally, new tests in Playwright and old ones ported by priority, while both suites run in CI until the old one is retired.

Question 48

How do you clean up test data created by automation?

Accepted Answer

Each test deletes what it created in an after hook, usually via API. For anything missed, a scheduled cleanup job removes records with our automation prefix, and lower environments are periodically refreshed.

Question 49

Your nightly suite has a 70% pass rate. How do you get it to 95%+?

Accepted Answer

I categorise failures over several runs into product bugs, flaky tests, data issues and environment problems, then attack the biggest bucket first. Flaky tests get quarantined and fixed, data gets isolated per test, and environment issues go to DevOps with evidence.

Question 50

Tell me about a critical bug your automation caught.

Accepted Answer

Answer with a concrete story: which test caught it, what the impact would have been, and what happened next. For example, a nightly checkout test caught a tax calculation error after a pricing service change, which was fixed before release.

Interview Questions

01Explain the automation framework used in your project.BeginnerFramework & Project

02How do you decide which test cases to automate?BeginnerStrategy & Planning

03What is your day-to-day work as an automation engineer?BeginnerFramework & Project

04How do you handle dynamic elements whose IDs change?BeginnerHandling UI Challenges

05How do you handle iframes and multiple windows?BeginnerHandling UI Challenges

06How do you automate file upload?BeginnerHandling UI Challenges

07What wait strategy do you follow in your framework?BeginnerHandling UI Challenges

08How do you handle flaky tests in your project?IntermediateExecution & Stability

09Your tests pass locally but fail in CI. How do you debug it?IntermediateCI/CD & Reporting

10How did you implement parallel execution?IntermediateExecution & Stability

11How do you manage test data in automation?IntermediateStrategy & Planning

12What reporting do you use and what does it include?IntermediateCI/CD & Reporting

13How are your tests integrated into CI/CD?IntermediateCI/CD & Reporting

14How do you automate flows with OTP or CAPTCHA?IntermediateHandling UI Challenges

15How do you reduce maintenance effort when the UI changes often?IntermediateFramework & Project