Full-stack engineer · Patna, India · Open to onsite / remote
Portfolio · Resume · LinkedIn · X · Email

I build SaaS products, developer tools, and small utilities end to end — mostly with TypeScript, Next.js, Postgres, and AI APIs. I'm self-taught, product-first, and looking for a full-stack or AI-product engineering role, onsite or remote.

I'm most useful on a small team that needs someone to take a feature from schema to shipped UI.

HackMeet — live 1:1 video rooms for developers. GitHub-verified, with a skip button when the conversation isn't working. WebRTC carries the media; Socket.IO handles signaling and matchmaking; the client and server share TypeScript types.

The beta is in development and the waitlist is open. Visit HackMeet →

BlogDrop — turns RSS feeds into one clean reading feed. Cheerio extracts each article; Gemini writes the summary, tags, and key takeaways; Inngest runs ingestion with retries and rate limits.
Next.js Postgres Drizzle Inngest Gemini · Code

QueryMate — ask a Postgres or MySQL database a question in plain English. It reads the schema, generates the query, and returns the result table in chat.
Next.js Postgres MySQL Drizzle Vercel AI SDK Groq · Code

SmartForm — a Chrome extension that reads form labels and context, asks Gemini or Groq for the values, and fills the fields. The user stays in control of the final submit.
TypeScript React WXT Gemini Groq · Code

Five live utilities: Temp Mail, Compressly, LowPDF, PDFAny, and Snapshot. LowPDF and PDFAny process files entirely in the browser — no upload, no server copy.

Frontend — TypeScript, React, Next.js (App Router), Tailwind, shadcn/ui, Zustand, TanStack Query
Backend — Node.js, Express, PostgreSQL, Drizzle, Better Auth, Inngest
AI & realtime — Gemini, Groq, Vercel AI SDK, Socket.IO, WebRTC
Tooling — Bun, GitHub Actions, Vercel, WXT, Arch Linux

Why I Built BlogDrop — the ingestion loop, the cleanup, and the parts that broke.
Next.js App Router in Practice — six months of production code: what held up and what I reverted.

I learned to code by shipping. Pick a problem, build the whole thing, put it online, then fix what real use exposes. That's still how I work — schema, API, UI, deploy.

I care about the unglamorous parts: clean data models, background jobs that can retry, error states, and pages that stay fast on a weak connection. I work from Arch Linux with Hyprland, mostly in Kitty, with Vim keys in VS Code.

Open to full-stack and AI-product roles, onsite or remote, and to interesting freelance builds.

hello@dhirajarya.in · Resume · LinkedIn · X