$ whoami → QA automation engineer, Tokyo

I find the bug before your users do.

Test automation across AI, automotive, telecom and e‑commerce. Playwright, Appium and K6, wired into CI that stays quiet unless something is really wrong, plus evaluation harnesses for AI assistants.

Get in touch Résumé (PDF)
  • 5+years in QA automation
  • 4industries tested
  • --:--Tokyo time (JST) right now

In Japan a card is offered with both hands. Click to turn it over.

$ ls work/

Four domains, four kinds of failure.

ai-ocr/
Cogent Labs · now

Accuracy is a spread, not a yes or no.

QA Automation Engineer on an AI and OCR product, promoted to P3 within the first year. Checks compare structured output with expected fields, and regressions show up between model versions instead of in production.

automotive-ivi/
L&T

Apps used while driving get no second chances.

In-vehicle infotainment testing on real devices. The interesting bugs live in interruptions: an incoming call, a lost connection, a screen that changes mid-task.

telecom/
L&T

Too many combinations to test equally.

Large device and configuration matrices call for choosing by risk. Cover the combinations customers actually use first, then widen the net.

e-commerce/
Earlier roles

Where a bug has a price tag.

Checkout, payment states and order flows. Browser automation guards the paths that carry revenue, so a release never has to ask whether buying still works.

$ plan --first-three-checks

Tell me what you ship. Here is what I test first.

    $ which --all

    Pick a tool. Here is what it does for me.

    $ cat how-i-work.txt

    How I work.

    A failing test explains itself.

    Names, traces and screenshots say what broke and where, so the person on call never has to guess.

    Test the risk, not the page count.

    Coverage numbers flatter. I start from what would hurt most if it broke and work outward.

    A suite nobody trusts gets ignored.

    Flaky checks get fixed or deleted. Green has to mean something, or the team stops looking at it.

    $ git log --author=ashish

    Experience.

    QA Automation EngineerCogent Labs · Tokyo · promoted P2 to P3 in the first year

    Test automation for an AI and OCR product. Structured-output checks, browser automation and CI that tells the team quickly when a model change moves the results.

    QA Engineer, automotive infotainmentL&T

    Testing an in-vehicle infotainment experience on real devices, with a focus on interruptions, reconnects and screens that have to stay readable at a glance.

    QA Engineer, telecomL&T

    Quality work across large device and configuration matrices, choosing what to cover first by how customers actually use the product.

    QA Engineer, e-commerceEarlier roles

    Checkout, payments and order flows, with browser automation on the paths that carry revenue.

    Computer Science degreeSRM University

    Computer science studies, the base for a career in test automation.

    Across roles Playwright (TypeScript) · Appium · Selenium · K6 · GitHub Actions · Claude Code

    $ mail --to ashish

    Something keeps breaking? Let's look at it together.

    Open to freelance QA projects and automation roles. Send a few lines about what you are shipping and where it hurts.

    LinkedIn Résumé (PDF)