Tools and experiments I build in the open, and products I ship on my own. I learn by shipping real software, not by collecting tutorials.
An open-source automation framework that brings Spring Boot-style conventions to Selenium — sensible defaults, dependency injection, and configuration over boilerplate, so teams write tests instead of plumbing.
A Model Context Protocol server that gives Claude and other AI assistants a real browser to drive — 84 tools for navigation, elements, assertions, and network control, with test-code generation for Java, Python, C#, and Gherkin.
An AI-powered test intelligence SaaS for Jenkins CI/CD — real-time build monitoring, flaky-test detection, and failure analytics, built multi-tenant from day one.
An AI-powered BDD toolkit for VS Code — one-click Cucumber/Behave test running with step-level results, natural-language scenario generation, and a quality linter, without leaving the feature file.
A self-hosted, real-time test reporting engine — a universal event model that any framework in any language can emit into, with live execution hierarchy, flakiness detection, and history in one place.
A full-stack test case management platform — plan, execute, and report on test runs with JIRA integration, bulk operations, and an automation API, built for QA teams rather than solo scripting.
A browser extension test automation toolkit — smart form filling, intelligent locator capture with health monitoring, and test recording that generates scripts for Playwright, Cypress, Selenium, and WebdriverIO.
A white-label SaaS for housing-society management — resident and payment tracking, expense management, and on-device AI financial insights, built offline-first with no cloud dependency.
A private personal wealth tracker for Android — net worth, investments, SIPs, fixed deposits and goals, with every record kept on the device and nothing sent anywhere.
Offline-first income and expense tracking for Android and iOS — multi-currency, five languages, recurring transactions, and a rules-based insights engine, with every record kept on the device.
Most of what's here is open source — I believe the best way to learn, and to give back, is to work in public, where the code, the mistakes, and the fixes are all visible. The rest are closed-source products, but built the same way: real tests, real architecture, nothing rushed.
See my work on GitHub ↗The tools I reach for — and a few I'm actively learning.
