Automation QA Engineer (Playwright, Selenium)

SHAMS FATHIMA

Automation QA Engineer Playwright · Selenium· Python · API Testing · AI-Assisted Testing Bangalore, India · +91-997**-***** · ******************@*****.*** linkedin.com/in/shams-fathima-b240771a0 · github.com/shams2666 PROFESSIONAL SUMMARY

Automation QA engineer with 4 years building and owning test automation for enterprise web platforms. Currently drive automation and manual test design for Apple's workplace real-estate platform: a 350-scenario Playwright suite across 38 spec files covering ergonomics assessments, space requests, reprographics, project intake, and workspace/headcount management, tested across admin, end-user, and super-admin roles. Automated 376 of 746 documented regression cases (~50% coverage); author the manual Radar suites behind that coverage as well (448 cases across 35 suites), reviewing specs and acceptance criteria before execution to catch requirement defects pre-build. Built a self-healing locator framework that cuts maintenance effort ~40% while holding 98% script reliability through UI churn; use Claude daily for test design, test-data generation, and failure triage.

CORE COMPETENCIES

•Test Automation: Playwright (JavaScript), Selenium WebDriver (Python), PyTest, Page Object Model, self-healing locators, reusable component libraries, parallel execution.

•Manual & Exploratory Testing: Test-case authoring and suite management in Radar, specification and acceptance-criteria review, requirement-defect escalation, exploratory testing, accessibility and keyboard-navigation validation, responsive-breakpoint testing, negative and boundary cases, empty-state coverage, defect documentation with screen-recording evidence.

•Quality Engineering: Test strategy, risk-based regression planning, end-to-end & functional testing, smoke & sanity gates, REST API testing, defect analysis & root-cause validation, release readiness sign-off.

•AI-Assisted QA: LLM-assisted test-case authoring from acceptance criteria, edge-case and test-data synthesis, automated failure triage and script debugging, coverage-gap analysis.

•Delivery & Collaboration: Agile Scrum, sprint planning, defect triage, production readiness reviews, cross-functional work with engineering and product, distributed/remote collaboration. QUALITY SYSTEMS SHIPPED

Playwright End-to-End Regression Suite: Apple Workplace Platform (Live)

•Problem: Ergonomics assessments, space requests, headcount reporting, and workspace-footprint workflows span multiple services and were regression-verified largely by hand each release: slow, inconsistent, and the main risk to release confidence.

•Built and maintain a Playwright suite of 350 end-to-end scenarios across 38 spec files, covering ergonomics assessment workflows, space requests, reprographics, project intake, and workspace/headcount management, each exercised across admin, end-user, and super-admin permission paths over the same shared UI.

•Cleared the enterprise-SSO barrier that blocks most UI automation: global setup performs a TOTP two-factor login once, merges and persists the authenticated session as storage state, and every spec replays it, turning a hard-to-automate auth wall into a one-time cost and unlocking 10-worker parallel execution.

•Structured on Page Object Model, with config-driven environment switching (dev/UAT) so the same specs run against any tier, encrypted credential handling, and Allure reporting for per-run failure evidence.

•Version-controlled in Git and executed through a CI pipeline, run on demand ahead of each release, written CI-ready; retry counts and worker allocation already key off the CI environment flag.

•Impact: Automated 376 of 746 documented regression cases (~50% coverage), raising admin-flow coverage from ~31% to ~56% in eight weeks, tracked case-by-case in a maintained coverage report. Triage failures and drive them to resolution with the development team through Apple Radar.

Manual Test Design & Specification Review: PlacesCentral Buildings (Live)

•Problem: Specs change mid-sprint and arrive internally inconsistent. Testing against a stale or contradictory spec produces invalid bugs, burns developer cycles, and erodes trust in QA.

•Authored 448 test cases across 35 Radar test suites covering DMS template administration, report builder, saved views, ACL/permission changes, and per-module column layouts, including 28 accessibility cases (20 keyboard-navigation), 15 responsive-breakpoint cases, and 29 empty-state cases.

•Review specifications before execution: on one Edit-Mode radar, compared three spec workbooks against the mockup and raised nine documented gaps and conflicts, then tracked each through a v2 respec and verified what was actually fixed. Shams Fathima · Automation QA Engineer Page 1

•Separate spec defects from implementation bugs: spec issues go back to the radar author as comments, only observed behavior is filed as a new radar. Gate execution on radar state so cases run against builds in verify, not while development is still in analyze-fix.

•Impact: Requirement defects caught before execution rather than surfacing as invalid bugs. Test plans, field-level checklists, and runbooks documented so any QA engineer can execute the suite unaided. Self-Healing Test Automation Framework (Open Source, 2025)

•Problem: UI suites break on cosmetic DOM changes; a large share of the maintenance budget goes to rewriting locators rather than finding real defects, and flaky failures erode trust in the suite.

•Built a Selenium/PyTest framework that intercepts element-lookup failures at the driver layer, inspects the live DOM, and generates candidate locators by scoring stable attributes (name, id, data-testid, placeholder, aria-label, visible text) against the original element, substituting the highest-confidence match so the run continues instead of failing.

•Separated into a healing driver wrapper, a candidate-scoring engine, a locator store that persists successful heals for future runs, and a dedicated healing log, layered over Page Object Model, with a demo harness whose page deliberately mutates a field's id to exercise recovery end-to-end via its stable name/placeholder attributes. Deliberately a deterministic heuristic, not a machine-learning claim: scoring rules are inspectable and every healing decision is auditable in the log.

•Impact: ~40% reduction in locator-maintenance effort and 98% script reliability sustained through UI churn. Open-sourced at github.com/shams2666/self-healing-selenium-framework. AI-Assisted Test Design & Triage Workflow

•Problem: Authoring cases from user stories and triaging red builds are the two slowest steps in a sprint; both bottleneck how much new functionality can be covered before release.

•Built a working loop with Claude: draft candidate test cases directly from acceptance criteria, generate edge-case and boundary test data, and debug failing automation scripts, with a review gate where every generated case is validated against the real requirement before it enters the suite.

•Impact: New stories get covered in-sprint rather than trailing a release behind, and defects surface earlier in the cycle, with no unverified AI output landing in the regression suite. Selenium Automation Suite (POM): Zurich Insurance

•Problem: Policy-management and user workflows were regression-tested manually across frequent release cycles, with test scripts that were expensive to maintain as the UI evolved.

•Built and scaled Selenium WebDriver (Python) suites on a Page Object Model architecture, separating page structure from test logic so UI changes touch one file instead of dozens.

•Validated policy management and user workflow functionality across multiple releases, and owned the smoke and regression gates that signed off production readiness.

OS Migration Validation: SVCT

•Designed and executed 100+ test cases validating workflow continuity through an OS migration, catching breakages pre-cutover and preventing production downtime.

PROFESSIONAL EXPERIENCE

Infosys Limited (Client: Apple)2025 – Present

Automation QA Engineer · Workplace & Real Estate Platform Bangalore, India

•Own UI and end-to-end Playwright automation for Apple's workplace real-estate platform; design test cases from business requirements and user stories, and run functional, regression, smoke, sanity, and E2E cycles.

•Author and maintain Radar test suites for the PlacesCentral Buildings module; review specifications and acceptance criteria before execution, escalating requirement defects pre-build.

•Log, track, and verify defects through Apple Radar, driving resolution with development teams; participate in sprint planning, defect triage, release validation, and production readiness.

•Introduced an AI-assisted workflow (Claude) for test authoring, debugging, and test-data generation (see Quality Systems Shipped above).

Capgemini Technology Services India 2022 – 2025

Test Engineer Bangalore, India

•Delivered Selenium/Python QA across client engagements during the 3-year tenure, including the Zurich Insurance policy-management suite and the SVCT OS-migration validation (see Quality Systems Shipped above).

•Ran functional, regression, and smoke validation as the pre-deployment quality gate; partnered with Agile teams to shorten delivery timelines without trading away release stability. Shams Fathima · Automation QA Engineer Page 2

TECHNICAL SKILLS

•Languages: Python, JavaScript (DOM traversal & browser debugging), TypeScript(Working proficiency).

•Automation & Testing: Playwright, Selenium WebDriver, PyTest, Page Object Model, REST API testing, cross-browser testing.

•AI Tooling: Claude / Claude Code for test design, debugging, and test-data generation; LLM-assisted coverage analysis.

•CI & Tooling: Maven, TestNG, GitHub Actions, Git, Allure reporting, Apple Radar (test-suite and case authoring, defect tracking), JIRA, Agile boards, Excel-based suite-to-radar-to-case traceability. EDUCATION

Bachelor of Engineering, Biomedical Engineering 2022 ACS University

Shams Fathima · Automation QA Engineer Page 3

Contact this candidate