How does Selenium WebDriver work under the hood?

Your test calls the language binding, which sends W3C WebDriver commands as HTTP requests to a browser driver (chromedriver, geckodriver), which controls the real browser. Selenium Manager downloads matching drivers automatically, and WebDriver BiDi adds a two-way channel for events.

Which locator strategies do you use in Selenium, and how do you choose?

Prefer stable, unique attributes: id or a dedicated data-testid via CSS. Use CSS selectors for most cases and XPath only when you need text matching or DOM traversal CSS can't do. Avoid absolute XPath and styling classes, and agree test IDs with developers.

Implicit vs explicit vs fluent waits in Selenium, and why not to mix them

An implicit wait makes every findElement poll for presence up to a timeout. Explicit waits (WebDriverWait) wait for a specific condition such as visibility or clickability. FluentWait is the configurable base with polling interval and ignored exceptions. Mixing implicit and explicit waits causes unpredictable timeouts.

How do you handle StaleElementReferenceException in Selenium?

It means the element reference points to a node that was removed or re-rendered. Re-locate the element after the DOM changes instead of caching WebElements, wait for the new state with WebDriverWait, and use Page Objects that find elements lazily.

How do you implement the Page Object Model, and why do many teams avoid PageFactory?

A Page Object wraps one page or component: private locators, public methods that express user intent, and no assertions about business rules. Many teams avoid PageFactory because its proxies cache poorly with dynamic DOMs, hide waits and add little over plain By locators.

How do you run Selenium tests in parallel safely?

Give every test its own WebDriver session (ThreadLocal or framework-managed per-test instances), never share static drivers or mutable state, make test data independent, configure parallelism in TestNG or JUnit 5, and run against Selenium Grid or containers sized to the thread count.

Explain Selenium Grid 4 architecture

Grid 4 splits the hub into components: the Router receives requests, the Session Queue holds new-session requests, the Distributor matches them to Nodes by capabilities, the Session Map records which Node owns each session, and an Event Bus connects them. It runs as standalone, hub-and-node or fully distributed.

Selenium Interview Questions and Answers

beginner questions

intermediate questions

advanced questions

What Selenium interviews cover at each level

Entry-level rounds check that you can use the API correctly: locator strategies, findElement versus findElements, and why Thread.sleep is a bad wait. Mid-level rounds focus on synchronisation and recovery: explicit and fluent waits, StaleElementReferenceException, frames, windows and alerts. Senior rounds become Selenium framework interview questions: how WebDriver talks to the browser, how you structure page objects, how you run tests in parallel without sharing a driver between threads, and how Selenium Grid 4 routes sessions.

Concepts to study first

WebDriver architecture: your code sends W3C WebDriver commands over HTTP to a driver (chromedriver, geckodriver), which controls the browser. Selenium Manager can now download the matching driver automatically.
Waits: explicit waits poll for a specific condition; implicit waits apply to every element lookup. The Selenium documentation warns against mixing them because the resulting timeouts are unpredictable.
Locator strategy: stable ids and data attributes first, CSS for structure, XPath when you need text or axes such as following-sibling.
Thread-safe parallel execution: one driver per thread, usually held in a ThreadLocal<WebDriver>, and no static driver fields.

A worked answer: stale elements

Asked "how do you handle StaleElementReferenceException?", a weak answer is "catch it and retry". A strong answer explains the cause first: the element reference points at a DOM node that was removed or re-rendered, often after an AJAX refresh. The fix is to locate the element again after the change, ideally by waiting for the change itself, for example wait.until(ExpectedConditions.stalenessOf(oldRow)) followed by a fresh findElement. Page objects that store By locators rather than WebElement fields avoid most stale references, which is one reason many teams avoid PageFactory. See the full answer on stale element handling.

Framework design questions

Selenium framework interview questions usually start with "walk me through your framework". Interviewers want layers with clear responsibilities: a driver factory that creates browsers from configuration, page objects that expose user actions (loginAs(user)) rather than raw elements, test classes that read like specifications, and test data and environment settings kept outside the code.

Be ready to explain how the same suite runs locally and on Selenium Grid 4 through RemoteWebDriver, how you capture screenshots and logs on failure, and how you keep the suite fast as it grows: run in parallel, move checks that do not need a browser to the API layer, and delete tests that duplicate coverage.

Practise beyond reading

The Selenium practice lab resolves your XPath and CSS locators against real HTML and checks your Java wait, frame and page-object code against pattern rules (it is not compiled). Pair it with the Selenium waits explained article and the Java interview questions, since most Selenium frameworks are written in Java.