back

work

Batteleground of my personal projects that I've built to learn new technologies.

> Built a WhatsApp-first, multi-shop grocery platform with a Go backend, admin dashboard, and custom payment module I designed from scratch.> The standout feature is per-shop payment orchestration supporting manual UPI and Razorpay with encrypted credentials, signed webhooks, and audit trails.

> Built a sub-10 MB Windows and Linux desktop app with snappy file conversion and ready-to-use tools using near-zero memory.> Owned the full product lifecycle—from idea, documentation, design, and prototyping to engineering and production delivery.> Its a fast, privacy-first workflow that feels nearly weightless while handling real file operations locally.

> Built Nabhya’s backend to power a trusted NGO–volunteer platform.> Learned to turn complex product requirements into secure, scalable, and reliable backend systems.> Created a personalized feed-ranking engine that matches volunteers with relevant opportunities in real time.

> Built CIWarden to reduce wasted CI compute by making GitHub Actions runs observable, controllable, and priority-aware.> I added queue controls, selective PR verification, failure analysis, and replay checks so teams spend runner time where it matters.> The payoff: fewer redundant runs, faster feedback, and a CI system that actively fights compute waste instead of quietly funding it.

> Built MentionBase to turn scattered X posts into a searchable, structured knowledge base.> Learned to design fault-tolerant, event-driven systems that handle retries, concurrency, and data integrity without drama.> Simple mention becomes a durable ingestion command that resolves and routes the referenced post automatically.

> Built a discoverable HTTP/JSON system that turns AI-agent API failures into structured, triage-ready evidence.> Learned to engineer protocol contracts, secure ingestion, privacy-aware scrubbing, and cautious agent behavior end to end.> Its sharpest trick: deterministic clustering collapses noisy, differently worded failures into one signal—without pretending volume equals truth.

> Built Rushout with Colyseus: a browser-based 2D party game supporting fast, chaotic online rooms and local multiplayer.> Learned to build authoritative real-time gameplay with 60Hz server simulation, compact input packets, client prediction, and server acknowledgements.> The standout: remote players stay smooth through snapshot interpolation while predicted movement hides network delay across 13-player arenas.

> Building a distributed order system to learn how real services coordinate under failure, not just happy-path demos.> Using Kafka, an outbox, retries, idempotency, and observability to keep orders reliable across payment, inventory, and notifications.> The goal: make every order traceable, recoverable, and production-ready when dependencies inevitably break.

> I built for devs, a keyboard-first cockpit for inspecting and managing local ports and processes.> I learned to turn OS-level chaos into a reliable, cross-platform workflow with safe automation and clear state.> Its standout feature is a live terminal UI that combines port discovery, process control, aliases, blueprints, health checks, and history in one cockpit.

> Built a full-stack web analytics platform that turns raw website activity into clean, interactive insights.> Learned how to design event pipelines, session tracking, visitor identity, authentication, and data-driven dashboards.> Its unique feature is a plug-and-play tracker that creates shareable public analytics dashboards from a single site ID.

> A minimalist badge generator and remixer where developers can create and customize beautiful badges for their projects to grab eyes as much as possible.> Built with better-t-stack and React Query.

> A CLI tool to help you clean up your npm packages by removing unused dependencies from your file system.> It analyzes your files with the regrex patterns to identify files that are not being used and removes them, helping you maintain a lean and efficient device.

> FormFuzz is an automated google form filler that uses AI to generate realistic responses based on the form questions.> Built with Google App Script, Node.js, and Gemini API, it streamlines the process of filling out lengthy forms by providing accurate and contextually relevant answers or generate a dummy data for surveys.

> End-to-end encrypted file sharing that disappears without a trace.> No accounts, no tracking, no compromises.> Written in NextJS and backed by supabase and appwrite for database and bucket storage.

> Create your own design follow and create beautiful contribution graphs.> Build a healthy habit of contributing daily.> Built with NextJS, Tailwind CSS, and Github OAuth.

> Bunch of ready to use templates to skip the hectic setting up dev environment.> Scaffold your next project with a single command.> Built with JavaScript and Node.js.