TypeScript for Testing

Install Node LTS and know what node, npm, and npx each do without looking it up.

Asked Screening/env-setup: 'walk me through setting up a Playwright TS project from zero' or a live 'npm init playwright@latest' run-through.

named in the job descriptions

Read package.json, add dev deps, and run/author npm scripts (the pip+venv+pyproject analog).

Asked 'How do you pin/install test deps?', 'what runs when someone types npm test?' — often a live repo walkthrough.

named in the job descriptions

Import/export symbols across files the way Playwright suites are structured.

Asked Live coding: organize a suite into modules; 'default vs named export difference?'

named in the job descriptions

Read a tsconfig and change the few knobs that matter, knowing Playwright only honors allowJs/baseUrl/paths/references.

Asked 'What does strict mode buy you?', 'how do import aliases work in your framework?'

asked in interviews, not named in any job description

Know professional repos lint+format automatically and CI blocks on lint errors.

Asked 'How do you keep code style consistent across the QA team?' (hygiene signal, rarely deep)

asked in interviews, not named in any job description

Run type-checking as its own CI step because Playwright runs tests even with type errors.

Asked 'Playwright ran my test even though it had a type error — why, and how do you enforce types?'

asked in interviews, not named in any job description

Annotate the 90% of everyday types you actually write in a suite.

Asked Fundamentals (~25% of TS interviews): 'annotate this function', 'what is any and why avoid it?'

named in the job descriptions

Declare the shape of test data and API payloads — the level POM + data factories use.

Asked Fundamentals + live coding: 'model this JSON as a type', 'what's a union type?'

named in the job descriptions

Give the crisp distinction that is near-guaranteed in a TS screen.

Asked Near-guaranteed direct question: 'interface vs type — when do you use which?'

asked in interviews, not named in any job description

Explain why unknown is the safe alternative to any for untyped inputs.

Asked Common screen: 'any vs unknown — which and why?'

asked in interviews, not named in any job description

Handle absent values the way strict mode forces you to in real TS code.

Asked Code review / live: 'this may be undefined — handle it', '? vs ! vs ?? difference?'

asked in interviews, not named in any job description

Constrain a value to a fixed set (roles, environments, test tags).

Asked 'enum vs string union?', 'how do you type environments?'

named in the job descriptions

Narrow a union to a concrete type so the compiler lets you use it.

Asked Advanced chunk: 'how do you safely handle an unknown response?', 'what's a type guard?'

asked in interviews, not named in any job description

Reuse existing types to build test-data variants without redefining shapes.

Asked Type-system chunk (~35%): 'how would you build flexible test data?', 'name a utility type you've used.'

asked in interviews, not named in any job description

Read the generics Playwright ships and write ONE reusable typed helper — not generic-heavy libraries.

Asked Expected direct question: 'what is a generic and when would you use one?' — answer + Promise<T>/helper example.

named in the job descriptions

Understand that a Promise is a future value and Promise<T> types what it resolves to.

Asked 'What is a Promise?', 'what does Promise<void> mean on this method?'

named in the job descriptions

The #1 Python-to-TS gap: async/await syntax and that async fns wrap returns in a Promise.

Asked Live coding + concept: 'why is this function async?', 'what does await do here?'

named in the job descriptions

Handle errors from awaited calls the same way you use try/except in Python.

Asked 'How do you handle an expected failure/timeout in a test?'

named in the job descriptions

The single highest-leverage async fact — turn the Python background into a strength by explaining WHY.

Asked Near-guaranteed: 'what happens if you forget await?', 'why must every Playwright call be awaited?'

asked in interviews, not named in any job description

Run independent async work concurrently and avoid the classic loop-await bug.

Asked 'How would you speed up test setup with several independent async calls?', 'why doesn't await work inside forEach?'

asked in interviews, not named in any job description

Structure a spec file with tests, groups, and setup/teardown hooks.

Asked Live coding baseline: 'write a test with setup', 'where does login go?'

named in the job descriptions

Read and edit the config you touch constantly (browser matrix, env, reporting, retries).

Asked 'How do you run cross-browser?', 'where do baseURL/retries/workers live?' — capgemini explicitly probes config tuning.

named in the job descriptions

The dominant framework pattern JDs demand: a typed POM class wrapping a page.

Asked Very common live task: 'write a Page Object for this screen', 'how do you structure your framework?'

named in the job descriptions

Use modern role/testid locators and auto-retrying web-first assertions — the 2026 baseline, not CSS/XPath sleeps.

Asked Heavy 2026 focus: 'locator strategy?', 'why getByRole over CSS?', 'how does auto-waiting remove sleeps?'

named in the job descriptions

Extend base test with typed fixtures (loginPage, apiContext) — the senior differentiator vs beginners who only use beforeEach.

Asked Senior probe: 'fixtures vs beforeEach?', 'how do you share auth/page objects across tests type-safely?'

named in the job descriptions

Test REST APIs in the same TS suite and type/validate the responses — JDs pair UI(TS)+API.

Asked 'How do you test APIs in Playwright?', 'how do you validate a response schema?'

named in the job descriptions

Wire the TS suite into CI with reporting and parallel sharding — a frequent hard must-have.

Asked Must-have: 'how do you run this in Jenkins/GitLab CI?', 'how do you shard/parallelize in CI?'

named in the job descriptions

Compose multiple fixture sets declaratively — highest single leverage 'framework-owner' signal.

Asked Architect-level: 'how do you compose fixtures across UI and API suites?'

asked in interviews, not named in any job description

Intercept and stub network calls to make complex scenarios deterministic — a 2026 interview staple the checklist underweighted.

Asked Common Playwright question: 'how do you mock an API response / test an error state without a real backend?'

named in the job descriptions

Log in once and reuse the session across hundreds of tests instead of re-logging-in — explicitly probed for senior SDETs in 2026.

Asked Senior probe: 'how do you handle auth across a large suite without logging in every test?'

asked in interviews, not named in any job description

Debug failures with the trace viewer — 2026 baseline; not knowing it reads as not knowing async in a backend interview.

Asked 'How do you debug a failing/flaky test?', 'what does the trace viewer show you?'

asked in interviews, not named in any job description

Author a reusable domain assertion — a senior/architect signal, name-drop level.

Asked Senior probe: 'have you extended expect / built custom matchers?'

asked in interviews, not named in any job description

Recognize that some TS UI roles use Cypress rather than Playwright.

Asked 'You've used Playwright — how does that transfer to Cypress?' on Cypress-shop roles.

named in the job descriptions

Recognize Cucumber/BDD-in-TS since several pool JDs list it as a plus or stack element.

Asked 'Have you used Cucumber/BDD?' — a fit filter on specific roles, not universal.

named in the job descriptions