{
"name": "Muskaan Timbadiya",
"day_job": "Senior Quality Engineer @ Cognizant",
"night_job": "Building AI tools with cultural relevance for India",
"based_in": "Kolkata, India",
"years_experience": 5,
"philosophy": "Quality is not an act, it is a habit. — Aristotle",
"currently_exploring": ["Agentic AI", "Multi-agent orchestration", "MCP servers"]
}By day, I design test strategy and automation frameworks that keep retail-scale software honest. By night — and at every hackathon I can find — I build GenAI products named in Hindi and Sanskrit, made for people who don't always see themselves reflected in AI products: exam-stressed students, home cooks, first-time voters, and soon, a stadium full of football fans.
Automation & Testing
AI / GenAI Stack
Languages, CI/CD & Tools
⚽ CrowdIQFIFA World Cup 2026™ Stadium Traffic & Fan Guidance AI Platform, a full-stack Event Transport & Crowd Logistics Command Center built with React, Three.js (WebGL), Vite, Express, Tailwind CSS, and the native Google GenAI SDK powered by Gemini 3.5 Flash.
|
AI mental-wellness companion for NEET, JEE, GATE & UPSC students. Five specialist agents, a StudentContextMCP server, mood tracking, an Agent Activity Feed, and a Vent Wall — built on Google ADK for a Kaggle AI Agents Hackathon.
|
🗳️ BallotBuddyMultilingual AI civic assistant helping Indian citizens navigate the 2026 election process — Hindi, Bengali, Tamil, Telugu & English, deployed on Cloud Run.
|
Multi-agent productivity assistant with an LLM router delegating to Task, Calendar & Notes agents, plus a metrics dashboard.
|
|
MCP-style agent + tool architecture serving live CoinGecko price data, containerized and deployed to the cloud.
|
|
Production-ready API testing framework — modular structure, env-based config, full CRUD coverage, CI/CD reporting.
|
End-to-end UI automation for booking & refund journeys with dynamic data, isolated runs, and auto-captured failure artifacts.
|
|
Risk-based testing models, release readiness checklists, metrics tracking, and governance templates for Agile teams. |
A custom MCP server that automatically repairs broken Playwright locators inside a CI healing loop, using real Chromium DOM snapshots and semantic locator strategy.
|
| Achievement | Impact |
|---|---|
| 🧪 Regression testing strategy | Identified 200+ critical issues, +25% user satisfaction |
| ⚡ Selenium → Playwright migration | Reduced testing duration by 30% |
| 🎯 UI automation suite | Consistent 80–90% pass rate |
| 📦 Test optimisation | Streamlined 900+ cases → 350 focused tests |
| 🔒 Defect management | Achieved 99.9% uptime |
Generated once a day by update-stats.yml using github-readme-stats-action and github-readme-streak-stats — both run inside this repo's own GitHub Action and commit static SVGs, so there's no live third-party endpoint to go down.
- ⚽ CrowdIQ — FIFA World Cup 2026™ Stadium Traffic & Fan Guidance AI Platform
- 🧘 Expanding MindMitra's multi-agent architecture post-Kaggle submission
- 🏛️ Managing accounts as Building Treasurer — with a proper web ledger and Excel exports
- 🌱 Deepening agentic AI patterns: MCP servers, multi-agent routing, tool-use design
- 🎤 Showing up at Kolkata's builder community — PromptWars, hackathons, demo days
I'm always up for a conversation about test automation, GenAI for underserved use cases, or the next hackathon idea.
