Amit Sheoran - Senior Lead QA | Test Automation & AI-First Quality Engineering

Currently executing

ElectreeFi. Charging infrastructure,
tested like infrastructure.

An IoT and eMobility platform for the Electric Vehicle domain - charging management, battery swapping, fleet management and EV business ERP. I architected its mobile test automation platform from scratch.

Full regression - Android, real device

345 cases executed - OnePlus 11R 5G - calibrated against the live build

VERIFIED

300

Passed

20

Failed

25

Blocked

12

Crashes found

304 test cases auto-catalogued straight from the product test-case workbook, so names, steps and expected results never drift from the source of truth.
A step engine executing 2,128 workbook steps across 211 charging cases, classifying each one as app action, verification, calculation or simulator interaction.
Appium 3 + UiAutomator2 on real Android devices, go-ios + WebDriverAgent on iOS - four app and platform combinations from one shared code base.
A Flask self-service App Runner so the team launches suites, manages test data and reads KPIs without waiting on me.
Automatic API capture that parses the app's HTTP client logs out of device logcat into structured records, with ready-to-run Postman export - API evidence for every UI test.
Crash and ANR detection after every single test: scan the log slice, fail the test, dismiss the dialog, restart the app.
Replaced a prior suite where 22 of 26 tests auto-passed without ever touching the device. Green means green now.

Appium 3UiAutomator2go-ios WebDriverAgentPythonFlask mitmproxyOCPP / OCPIRazorpay Selenium (CMS)

About

The unglamorous part
decides the release.

I started as a tester at Askmebazaar in 2015, broke things at Healthians, automated negotiation systems at PrivateDeal, owned a 600-case framework and release validation at Airlinq, and today lead quality for ElectreeFi at TechPerspect.

What I do best is the part nobody puts in a demo: reading a logcat slice at 2AM, writing the assertion that catches the silent migration bug, building the database check that exposes what the UI was hiding, and pushing back on a "ship it anyway" PR with calm, specific evidence.

I treat automation as a software product, not a pile of scripts. The suites matter, but so do the runner, the dashboard, the evidence pipeline and the report - because those are what a team actually uses on a Friday evening before a release.

The last two years changed how I work. I stopped using AI to draft test cases and started giving it skills and tools - an authored skills library, MCP connectors wired into Jira and the database, and coding agents that refactor my framework while I review the diff. The job is still judgement. The typing is no longer the bottleneck.

How I work

EVIDENCE

A result nobody can audit is not a result

Every test carries before and after screenshots, the API calls it made, and its slice of the device log. When I say it passed, you can check.

DATA LAYER

Assert where the truth lives

The UI is the last place a bug shows up and the easiest place to be fooled. I query the database, read the logs, and check the API response - then look at the screen.

HONESTY

Green has to mean green

I inherited a suite where 22 of 26 tests passed without ever touching the device. A dashboard that lies is worse than no dashboard, so I rebuilt it on real assertions.

LEVERAGE

Build the tool, not just the test

If I am the only one who can run the suite, I am a bottleneck. So I ship runners and dashboards that let QA, DevOps and product launch their own runs.

AMIT SHEORAN · AVAILABLE

Role

Senior Lead - QA

Company

TechPerspect Software (ElectreeFi)

Experience

11+ years - 8+ in automation

Base

Gurgaon, India - open to hybrid

Domains

EV & IoT - Telecom - Healthcare - E-Commerce

Core stack

Appium - Selenium - Java - Python - SQL

Open to

Senior QA - SDET - Test Lead roles

GitHub LinkedIn Medium Contact

Experience

Five companies. Eleven years.
Zero rollbacks.

Every role below expands. Open one to see what actually got built.

Architected the Appium mobile automation platform from scratch for two EV charging apps across Android and iOS - 304 catalogued cases, a 2,128-step engine, and a Flask self-service runner.
Built device-log API capture with Postman export and crash and ANR detection on every test, so every UI result carries API and stability evidence.
Own test strategy, QA process definition, defect governance and release readiness for cloud services and IoT-connected device workflows.
Automated CMS validation with Selenium and payment journeys including Razorpay wallet recharge with end-to-end balance verification.
Keep an honest scope line with stakeholders: what is calibrated on real devices versus what still awaits first-run calibration. Coverage reporting you can trust.
Mentor QA engineers on automation standards, locator and wait discipline, failure triage and code review.

AppiumAndroid + iOSPython FlaskSeleniumOCPP / OCPI Test StrategyMentoring

Designed and solely owned an end-to-end UI and API automation framework of 600+ test cases across 50 suites, running multi-client, multi-environment and multi-browser from a single codebase - 4 environments, 2 enterprise clients, 6 browser targets.
Built a custom Python test runner and a self-service dashboard indexing 300 test cases across 15 feature areas, adopted by QA, DevOps and delivery teams.
Created a database validation layer in Python and SQL over secure SSH tunnels asserting subscription, entitlement and billing state - surfacing defects the user interface concealed.
Stabilised a critical delivery-drop regression from 70 pass and 5 fail to 95 pass and 0 code-level failures across 100 cases, and published a 7-category root-cause catalogue the team still uses for triage.
Owned release validation across Dev, QE, Delivery and Client teams - 15 to 16 releases per cycle, 20+ production releases with zero rollback incidents.
Delivered onsite at Maxis, Malaysia (Dec 2023 to Feb 2024) supporting client UAT and go-live.
Introduced AI-assisted QA across the team, cutting manual regression effort by around 40 percent.

Robot FrameworkPythonSelenium JenkinsSQL / MySQLAnsible KubernetesRHEL / CentOSJIRA + Xray

Built a Java and Selenium WebDriver framework with TestNG and Maven from scratch - sign-up, login, booking, offer accept and reject, conditional counter-offers and trial expiry. Around 70 percent less manual regression effort.
Validated the Merchant API across three integration methods (iFrame, REST API, on-platform negotiation): payloads, authentication, pricing contract exchange and rate limiting.
Tested a ChatGPT-powered dynamic messaging engine where responses were generated live from user behaviour - my first real encounter with testing non-deterministic output against rules and thresholds rather than fixed strings.
Validated a security-focused password policy module: reuse prevention, policy invalidation, minimum length and multi-factor authentication for B2B and B2C users.
Best QE Award, January 2022.

JavaSeleniumTestNG MavenPostmanMySQLJMeter

Led QA for mobile and web releases across Android, iOS and web. Promoted to Team Lead in January 2019, mentoring the QA team and reviewing their test design.
Led validation of the H-Cash e-wallet and cashback engine: balance correctness, discount upper limits, eligibility calculated on amount spent, fulfilment-gated credit within 48 hours, non-transferability.
Validated a rules-driven report auto-verification engine gated on data availability, biological reference ranges, restricted parameters and payment state.
Tested WhatsApp Business API integration for transactional alerts - event triggers, mandatory opt-in and SMS fallback.
Backend validation with SQL, log monitoring on AWS CloudWatch and S3, crash analytics through Firebase and Fabric.
Employee of the Month (Jan 2020) and Mindful QE Award (2018).

Android / iOSSeleniumSQL AWS CloudWatchFirebaseJMeterTeam Lead

Functional, regression and system testing across seller onboarding, automated seller assignment, workflow routing, SLA and escalation rules and buyer and seller rating.
API testing and backend data verification with SQL across order, seller and commission tables.
Designed and executed JMeter load and performance tests on localized commerce transactions to find throughput bottlenecks.

Manual QASQLJMeterAPI

AI-first engineering

I don't prompt an AI for test cases.
I give it skills and tools.

Most "AI in QA" means pasting a requirement into a chat window. Mine means authored skills, connected tooling, and an agent that can act across the stack instead of only producing text.

An open-source QA Skills library

I wrote and published 20+ production-ready QA Skills for AI coding agents - Claude Code, Cursor and Codex. Each one is a self-contained SKILL.md teaching an agent to do exactly one QA task well, with a defined workflow, a strict output format and a self-review checklist. One responsibility per skill.

test-case-generatorrequirement-analysistest-plan test-strategyexploratory-testingjira-bug-writer root-cause-analysisbug-triagesmoke-testing regression-testingselenium-javaplaywright cypressrobot-frameworkappium api-testingsql-testingcontract-testing performance-testingaccessibility-testingsecurity-testing

github.com/amitsheoran/automation-qa-skills

MCP connectors in the loop

Model Context Protocol connectors wired into the QA workflow, so the agent reaches the tools instead of describing them.

jiraA test fails, the bug is raised automatically - with reproduction steps and evidence already attached.

supabaseDatabase reads for test data and backend state verification.

vercelDeployment context pulled straight into the run.

Day to day I run an agentic setup with Claude Code and Cursor, prompt engineering tuned per model, and AI-assisted failure analysis that separates genuine product defects from environment noise. Measured outcome: around 40 percent less manual regression effort.

Toolkit

Instruments, all battle-tested.

Highlighted items are what I am in daily, right now.

Mobile automation

Appium 3
UiAutomator2
go-ios
WebDriverAgent
ADB / logcat
mitmproxy
Android Studio
Real devices
Emulators

Web and UI automation

Selenium WebDriver
TestNG
Maven
Page Object Model
Robot Framework
ACCELQ
Cross-browser
Headless

API and data

Postman
REST Assured
Swagger
SoapUI
SQL / MySQL
REST and SOAP
OAuth / SSO
pymysql

Languages

Python
Java
JavaScript
SQL
Bash
C / C++

AI and agents

Claude Code
MCP connectors
Cursor
ChatGPT / OpenAI
Gemini
Skill authoring
Prompt engineering

CI/CD and infra

Jenkins
Git / GitHub
Ansible
AWS
Linux (RHEL, CentOS)
Kubernetes
Flask tooling

Process

Test strategy
Test planning
Defect governance
Release readiness
Agile Scrum
Shift-left
Mentoring

Performance

Apache JMeter
LoadRunner
Load / stress design
Bottleneck analysis

Selected work

Things I built that teams use.

TechPerspect - ElectreeFi

EV Mobile Automation Platform

Appium-based Android and iOS automation for two EV charging apps. Workbook-driven catalogue, step engine, Flask self-service runner, device-log API capture with Postman export, crash and ANR detection on every test.

304 cases - 2,128 steps - 4 app and platform targets

Open source

automation-qa-skills

A published library of QA Skills for AI coding agents. One responsibility per skill, each with a workflow, strict output format and self-review checklist - from test design and bug triage to Selenium, Playwright, Appium, API, SQL and security testing.

20+ skills - Claude Code - Cursor - Codex

Airlinq

Carrier Marketplace Framework

UI and API automation across 50 suites, multi-client and multi-environment from one codebase, with SQL backend validation, automated defect reporting and JIRA and Xray integration.

600+ cases - 4 environments - 6 browsers

Airlinq and TechPerspect

Self-Service Runner Dashboards

Internal web tools that expose automated suites to QA, DevOps, product and delivery teams. People launch their own runs and seed their own environments - the automation team stops being the bottleneck.

300 cases indexed - 15 feature areas

Airlinq

Acceptance Evidence Pipeline

Chrome DevTools Protocol full-page capture with auto-generated evidence documents, built to a global carrier's compliance and audit standard. Before and after frames per action, highlighted assertions, no masked errors.

Accepted as formal carrier acceptance evidence

PrivateDeal

Java Selenium Regression Suite

Java, Selenium WebDriver, TestNG and Maven framework built from scratch for booking, negotiation, counter-offer and trial-expiry journeys on a European hospitality platform.

Around 70 percent reduction in manual regression effort

Writing

Notes from the field.

Published on Medium. Mostly about AI, deployments and the things that break at 2AM.

01

Praise

What people I've shipped with say.

Unedited LinkedIn recommendations from managers, mentors and peers - 2018 through 2025.

I had the privilege of working with Amit Sheoran, a highly skilled and dedicated Senior QA at PrivateDeal. His expertise in custom dashboards, Shopify integration, API security testing and custom messaging using ChatGPT is unparalleled. Amit's proactive approach and attention to detail consistently elevate project outcomes. A true asset.

ASAtinder Singh
Lead Engineer - Laravel, VueJS, Node, AWS
Jan 2025 - worked on the same team

Amit is one of the most budding and enthusiastic people I have met in technology recently. His penchant for digging deep into issues and scenarios is fascinating - something you'll want in every team member. He doesn't just push himself further but influences the team to push the envelope.

SSShammi Shailaj
Web Tech Leader - AI, Go and Scalability
Jan 2021 - managed Amit directly

Amit is a fantastic person to work with and is not only a multi-skilled and insightful colleague but also an inspiring strategist. Very good person and great employee with very strong problem-solving skills. Wherever he goes, I'm sure he'll add great value to the company.

MSMehak Sharma
Strategic Partnerships at Yubi MENA - ex-PayU
Dec 2019 - different companies

Amit Sheoran is among the rare breed of hard-working professionals that are an asset to any team. He is focused on his work as a software test engineer and serves as a mentor to his peers whenever required. End-to-end knowledge of the product area - a crucial resource for the technology team.

AGAbhinav Gade
Product Leader - B2C, e-Commerce, Insurtech
Oct 2018 - was Amit's mentor

Amit Sheoran is a competent and detailed Software Test Engineer. His inputs are highly sought for detailed technical information - many come to him for clarification on work-related issues. Knowledgeable, dedicated and most of all helpful. An absolute pleasure to work with.

SVSonali Verma
Product Marketing and Growth - Human-Centered GTM
Oct 2018 - worked on different teams

Amit Sheoran is one of the best among all people I have ever met and is a loyal colleague who understands complex matters. A strong and goal-oriented team player - with every problem, there was a solution. Highly recommended.

VSVijay Prakash Saini
Senior ML Engineer - Quantiphi Analytics
Feb 2018 - studied together

Amit Sheoran is one of the hardest working people I know. He's thorough and usually goes beyond the scope of what is asked. He often comes up with relevant use cases not previously considered. In his QA position, he gets joy in breaking things, quoting that it's a great stress relief.

NMNikhil Malik
Web, Scalability, Security, Impactful Solutions
Feb 2018 - worked on the same team

7 recommendations - 2018 to 2025 - See all on LinkedIn

Education and certification

Where it began.

2011 - 2015

B.Tech, Electrical and Electronics Engineering

B.K. Birla Institute of Engineering and Technology, Pilani, Rajasthan - Full time

2011

Higher Secondary School

Birla School Pilani - CBSE Board

2009

Secondary School

Birla School Pilani - CBSE Board

Selenium Essential Training - LinkedIn Advanced JMeter - LinkedIn Postman Essential Training - LinkedIn DevOps Foundations: CI/CD - LinkedIn Atlassian Agile Project Management Professional

Onsite Client Recognition - Maxis Malaysia 2023-24 Best QE Award - PrivateDeal 2022 Employee of the Month - Healthians 2020 Mindful QE Award - Healthians 2018

Get in touch

Let's ship something
that doesn't break.

Open to senior QA, SDET and test leadership roles, and to consulting on test automation, mobile QA and release engineering. Fastest reply is on WhatsApp.

Email

option.amit@gmail.com

Phone / WhatsApp

+91 88244 44866

Location

Gurgaon, India
Open to hybrid - Delhi NCR

Response time

Usually within 4 hours

GitHub LinkedIn Medium Twitter

ElectreeFi. Charging infrastructure,tested like infrastructure.