I'm Harish — a quality engineer who believes testing is a design discipline, not a final checkpoint. I build automation that teams actually trust, wire it into every release, and keep asking the uncomfortable question: what happens when this breaks?
Today I test a large-scale EV-charging platform at Shell Recharge Solutions end to end — web, mobile app, APIs, payments, partner integrations and the data behind them.
Before that: fintech lending flows at Moolya, mobile and web at Glynk, and games across five platforms at Indium. Lately I build AI tools that write and heal their own Playwright tests.
0+
Years in quality engineering
0+
Test tools & frameworks used in production
0
Platforms tested — web, iOS, Android, WebGL
0
Industries — EV charging, fintech and gaming
02 Expertise
What I bringto a team.
01 — AIRare in QA
“Add an item to the cart, check out and verify the order total.”
I don't just use AI tools — I build them. My own platform turns plain-English test cases into Playwright scripts, finds locators from the live DOM and repairs tests when the UI changes.
Your team getsMore coverage without more headcount, and tests that keep up when the UI changes.
0Brittle selectors — locators come from real interactions, not text guesses
Claude / OpenAIPlaywrightBrowserUseGeminiPixelmatchCopilot
02 — CodeDeveloper fluency
Start session
POST /sessions · 201 · 142ms✓ ui.spec passed
Full-stack fluency
I've designed and coded my own products end to end — frontend, APIs and databases. So I test with an understanding of how the system is built, not just how it looks.
Your team getsA tester who can read your pull request, reproduce bugs at code level and talk to developers in their language.
UI→DBComfortable across the whole stack I test — frontend, APIs, queues and databases
FastAPINext.jsReactTypeScriptPostgresPrisma
03 — DepthBelow the UI
Below the UI
A green screen doesn't mean the system is right. I follow every request below the UI — API responses, application logs, event streams and database records — to prove the data is correct end to end, including across third-party integrations.
Your team getsIntegration bugs caught where they start — in messages and data — before customers or partners see them.
30+External partner integrations tested end to end — logs, events and data checked on both sides
synthetic checks on productionalert on first failure
Quality after release
Testing doesn't stop at deploy. I design synthetic checks and monitors that keep testing live systems and raise an alert the moment something breaks in production.
Your team getsProblems caught in production before users — or clients — notice them.
24/7Playwright monitors that watch live sites and alert on Slack, WhatsApp or email
Working on my own product ideas — business cases, unit economics, customer conversations — taught me to prioritise testing by business impact, not by test-case count.
Your team getsTesting effort spent where it protects revenue and user trust, instead of spread evenly.
ROIRisk-based plans that put revenue-critical flows first
I join refinement and design reviews, question requirements early and help write acceptance criteria — so bugs get prevented, not just found.
02
Bug reports devs like
Every defect comes with clear steps, expected vs actual, logs, API traces and database evidence. Less back-and-forth, faster fixes.
03
Own the release
Test plans, sign-off and post-production checks are mine. Stakeholders get a plain-language report: what was tested, what's at risk, go or no-go.
04
Work across teams
From developers and product owners to 30+ external partner companies — I'm comfortable leading integration testing with people outside the company.
05
Automate the boring
If a check runs twice by hand, it's a candidate for automation. Manual time goes to exploratory testing, where humans find what scripts miss.
06
Keep learning
Game QA to fintech to EV charging to AI agents — I pick up new domains quickly and bring fresh tools back to the team.
03 Experience
From test casesto quality engineering.
2020Manual testingGames & mobile apps
2021AutomationSelenium, Java, regression suites
2023SDETAPI, mobile & fintech automation
2024 — nowQuality engineeringE2E, backend, CI/CD & AI
Next chapter
Design and run end-to-end tests of EV-charging workflows on simulators and live charging sessions.
Functional and API testing of OCPP 1.6 (JSON) and OCPI 2.1.1 / 2.2.1 between eMSP and CPO systems.
Led onboarding and integration testing for 30+ roaming partners with internal and external teams.
Extended UI and API automation frameworks and wired them into TeamCity for continuous regression.
Validated Kafka message flows, Dynatrace logs and MySQL data for sessions, CDRs, locations and tariffs.
Payment testing across Payter, RFID and mobile-app sessions; validated OneMPS pricing and tariff distribution.
DomainEV charging
FocusE2E · API · Protocols
StackKafka · TeamCity · MySQL
End-to-end testing of loan workflows — PAN/Aadhaar validation, risk checks and loan crediting.
Validated onboarding journeys plus deposit, bill-payment and credit-card flows.
Automated critical flows with Python, Pytest and Appium on Android.
Monitored AWS logs, ran backend validations and joined bug-bash sessions.
DomainFintech
FocusMobile · API
StackPython · Appium · AWS
Smoke, sanity, functional and exploratory testing to keep releases stable.
Automated regression suites in Selenium with Java.
Mobile testing on iOS and Android; cross-browser checks on Chrome, Firefox and Safari.
DomainWeb & mobile apps
FocusWeb · Mobile
StackSelenium · Java
Started in game QA, testing across iOS, Android, WebGL and Kindle and working directly with game developers.
Moved into product testing — test scenarios from functional specs; functional, integration and regression cycles in Jira.
First steps into automation with Selenium and performance testing with JMeter.
RolesTest Associate → Test Engineer
FocusManual · Mobile · Load
StackSelenium · JMeter · Jira
04 Selected work
Tools I builtto test better.
Built at workInternal toolFull-stack
EV roaming test bench
A full-stack QA dashboard that simulates EV chargers and validates roaming APIs end to end — so the team can test a complete charging session without a physical charger.
01Built-in OCPP charger simulator over WebSockets — boot, heartbeat, authorize, remote start/stop and meter values.
02One-click orchestrator runs a full session: boot → start (remote, RFID or Plug & Charge) → meter values → stop, with a live timeline.
03OCPI response validator checks live APIs against the spec — type, enum, format, length — with field-level diagnostics.
04Live traffic log per charger, encrypted token caching, and multi-region / multi-version switching.
Outside work I like building things to understand problems from the inside. Not all of them shipped — but each one sharpened how I think about quality: question the assumption, find the weak point, then test there.
Side projectWhat it taught me
01AI + observability
Trizhul
AI root-cause analysis for logs
Problem: when production breaks, engineers dig through logs, metrics and traces by hand. Explored: an AI layer you can ask "why did this fail?"
Root cause is a testing skill, not just a developer one.
→
In my testingWhen a test fails, I correlate logs, metrics and traces — so my bug reports point to the cause, not just the symptom.
02Full-stack build
Hyperlocal
Neighbourhood commerce & delivery
Problem: local shops can't compete with quick-commerce apps. Explored: a customer app, rider app and shop portal on one shared backend.
Systems break where they connect.
→
In my testingI now aim end-to-end tests at the seams — between apps, services and third parties — where the real bugs hide.
03Market research
Footfall
Marketing autopilot for local businesses
Problem: salons, clinics and cloud kitchens can't tell which marketing actually works. Explored: a platform that plans the campaign, makes the creative with AI and measures the result.
Opinions aren't evidence — behaviour is.
→
In my testingI validate against real user behaviour and production data — not just the spec, or a "looks fine" in a demo.
04Product design
AdWear
Wearable digital advertising
Problem: outdoor ads have no targeting or measurement. Explored: a jacket with a video screen, plus a dashboard for brands to book screen time by location.
Every system has a weakest link.
→
In my testingI map dependencies first and test the riskiest component early — before it can take the whole release down.
06 Toolbox
The kit I reach for.
Automation 04
Playwright
Selenium
Cypress
Appium
Frameworks 06
CodeceptJS
BDD Cucumber
TestNG
Pytest
Page Object Model
Hybrid
Languages 05
Java
JavaScript
TypeScript
Python
SQL
API & performance 04
Postman
Swagger
JMeter
REST APIs
CI/CD & observability 08
TeamCity
GitHub
Dynatrace
OpenSearch
Kafka UI
MySQL
AWS
TestRail · Jira
AI-assisted 05
Claude / OpenAI
BrowserUse agents
Gemini 2.0 Flash
GitHub Copilot
Pixelmatch
— Education
Bachelor of Engineering
BTL Institute of Technology & Management, Bangalore