Q&A · Straight answers
What people ask first.
Short, checkable answers — the same ones I would give in a first message, and the same ones an answer engine can quote.
What people ask first.
Short, checkable answers — the same ones I would give in a first message.
A Software Engineer (QA) based in Ahmedabad, India, with 7 years 4 months in QA since June 2019. He has been at Openxcell since April 2023, testing the Speed crypto payment platform, and before that spent nearly four years at Auxano Global Services. He also built and maintains Aegis-QA, an internal AI-driven test automation platform.
Openxcell, in Ahmedabad, since April 2023, as a Software Engineer (QA) on the Speed crypto payment platform. Some directory listings still show the earlier role, QA Engineer at Auxano Global Services, which ended in April 2023, along with an experience figure that stopped counting at the same time. The current figure is 7 years 4 months of QA, from June 2019.
Both manual and automation, which is deliberate rather than transitional. He designs the test cases, automates the ones worth automating, verifies the result in the database rather than in the UI response, and load-tests the paths where volume is the failure mode. Payments is the domain he knows best.
Money movement on crypto rails. On-chain BTC, ETH, USDT (ERC-20) and XAUT, plus the Lightning Network. Each has its own confirmation timing and decimal rules. ERC-20 smart contract calls, including ABI encoding, revert paths, emitted events and access control. KYB and KYC onboarding through a third-party identity vendor, both LLC and Corporation structures. And ledger reconciliation, which is the class of bug where the UI reports success and the database disagrees.
Playwright and Pytest for web, Appium for Android, Postman and Newman for API and contract testing, JMeter for load, SQL for database verification, and Datadog for tracing a failed request. Python, JavaScript, TypeScript and SQL. Docker, AWS EC2 and CI pipelines for the plumbing. Claude, Claude Code, Cursor and ChatGPT as daily working tools, not novelties.
An internal QA platform he built and maintains. It records a real session on the real application, compiles it to readable Gherkin and a runnable Playwright spec, runs UI, API, flow and security layers across five runner modes, and proposes selector repairs that are verified in a sandbox and never applied without a human approving them. Failed runs are classified and diagnosed automatically. Python 3.12, Playwright, asyncio, FastAPI and Next.js.
Merchant checkout, payment links, refunds, crypto payouts and instant payouts, transfers, swaps, cashback, wallet address verification and audit trails, across BTC, Lightning, ETH, USDT and XAUT. Plus WooCommerce and e-commerce plugin integrations, and checkout localization across devices.
Open to work, remote or hybrid, from Ahmedabad, Gujarat, India. He is interested in QA engineering and test-automation roles, especially anything with payments in it. Reachable at malayadalja123@gmail.com.
QA, by profession and by choice. AI is the thing he went and learned because the tooling he needed did not exist, and Aegis-QA is the result: a QA platform built with AI rather than bought from a vendor. Claude, Claude Code, Cursor and ChatGPT are daily working tools for him, used for test design, failure triage, scripting and reading unfamiliar code. The QA judgement is his; AI is how one person shipped a platform that would normally take a team.
Because the second half of this job is tooling. Seven payment rails with five assertions each, on every release, does not scale by hand. So the useful question is not whether someone can write a test case but whether they can build the thing that runs ten thousand of them and tells you which one matters. This site runs its own accessibility and performance suite against itself, in the visitor’s browser, as a small demonstration of exactly that.