I build automation frameworks that catch real bugs — not just flag value mismatches.
10+ years in QA. Now integrating LLMs into automation frameworks to solve problems that brittle assertions can't.

Most automation engineers write tests. I build the infrastructure underneath — frameworks that are maintainable at scale, CI pipelines that don't flake in production, and increasingly, systems that use LLMs to validate what "correct" actually means for a given API response.

Three things I'm focused on right now:

• LLM-based semantic validation — replacing brittle exact-value assertions with AI reasoning that understands intent
• Flaky test elimination — locator strategy, async handling, and test isolation done properly
• AI-augmented QA infrastructure — self-healing locators, intelligent failure triage, LLM-assisted test generation

→ View Repository

The core problem: traditional assertions fail the moment a value changes by one character, even if the response is semantically correct. This framework replaces that with LLM-based reasoning.

How it works:

• Sends API response + validation intent to an LLM
• Gets back: PASS/FAIL verdict + natural language reasoning + confidence score
• Falls back gracefully to schema validation when AI is unavailable

What's inside:

Key features:

• 🧠 AI verdict with reasoning + confidence score per test
• 🔁 Retry logic, response caching, rate-limit handling
• ⚙️ Config-driven: run in mock mode without API keys
• 🧩 Modular — swap the LLM provider without touching test logic
• 📊 Validated against real-world APIs (ReqRes, REST Countries)

Traditional automation checks exact values. This framework validates correctness intelligently — understanding what a response should mean, not just what it should equal.

→ View Repository

Production-ready Cypress framework for UI and API testing — built to address real-world challenges like flaky selectors, brittle flows, and hard-to-maintain suites.

What's inside:

• Page Object Model with centralized locator strategy (data-testid first)
• UI + API automation in a single framework
• Multi-environment support (dev / staging / prod) via config flags
• CI/CD wired with GitHub Actions — runs on push and PR
• Mochawesome HTML reporting + failure screenshots

Stack: Cypress JavaScript GitHub Actions Mochawesome POM

→ View Repository

BDD-first Playwright framework combining Gherkin syntax with modern Playwright capabilities — built for teams that need readable, stakeholder-friendly test scenarios alongside robust automation.

What's inside:

• Gherkin .feature files for UI and API scenarios
• Multi-browser coverage: Chromium, Firefox, Safari
• Multi-environment config (dev / qa / uat)
• Triple reporting: Allure + Playwright HTML + Cucumber reports
• OCR integration via Tesseract.js for visual/text validation
• Parallel execution with video recording on failure

Stack: Playwright TypeScript Cucumber/Gherkin Allure BDD

Test Automation Cypress Playwright Selenium WebdriverIO

Languages JavaScript TypeScript Java

API & Performance REST Assured Postman Newman JMeter

BDD & Reporting Cucumber/Gherkin Allure Mochawesome

CI/CD & DevOps GitHub Actions Jenkins Docker

AI / LLM Integration OpenAI API Claude API LangChain Semantic Validation

Domain Experience Fintech Telecom Insurance SaaS

• Building production-grade LLM validation layers for API test suites
• Self-healing locator strategies using AI-assisted element resolution
• Integrating LLM-based failure triage into CI/CD pipelines
• Exploring RAG-based test knowledge bases for large codebases

Remote-first roles with Series A–C product teams moving fast:

• Senior SDET / QA Automation Architect
• AI + Testing focused engineering roles
• Freelance / contract engagements (available now)

Timezone: IST (UTC+5:30) — overlap-friendly with EU mornings and US afternoons.

The future of testing isn't writing more assertions —
it's building systems intelligent enough to understand what correct actually means.