Hey, I'm Raghav

Raghvendra Dwivedi.
Full-Stack Developer.

I build performant backends, real-time systems, and polished web apps — obsessing over the details that make software actually feel good.

see my work

open to opportunities

building

Crontab & Drop It

learning

System Design • DevOps • AI/ML • WebRTC

about

Hey, I'm Raghvendra — most people call me Raghav. I'm a full-stack developer specializing in backend systems, real-time applications, and performance optimization. I build with React, Node.js, and MongoDB, and I've shipped systems that are 30% faster than their predecessors.

When I'm not shipping code, I'm writing about it — 300+ technical articles and counting, reaching 10k+ monthly readers. I believe good engineering is invisible: it just works, fast, and never makes the user think twice.

I don't ship junk. I optimize queries, cache aggressively, and build things that hold up under real traffic. My best work happens when curiosity meets a hard deadline.

Developer Snapshot

Building real-time systems.
Optimizing everything.
300+ technical articles.
Hackathon-tested, production-ready.

where I've worked

Cangra Talents

Full-Stack Developer Intern

Integrated legacy PHP systems with modern Express.js APIs. Developed an AI-driven interview report generation system. Built 20+ responsive pages and dashboard modules with React & Tailwind CSS.

Apr 2025 — Present

Cangra Talents

General Content Strategist

Authored 300+ articles about an Interview-as-a-Service platform. Published blogs reaching 10,000+ monthly readers across recruitment, career guidance, and software development topics. Improved SEO and organic traffic by 40%.

Sep 2024 — Apr 2025

Self-Employed

Freelance Full-Stack Developer

Designed and developed custom web applications using MERN stack and Next.js. Optimized database queries and caching strategies to achieve 30% faster API response times. Implemented real-time features with WebSockets.

Jan 2023 — Present

things I've built

featured

NoteBook AI

2025

An AI-powered RAG assistant that lets you chat with your PDFs. Upload a document, ask questions, and get cited answers with source tracking — like having a research partner who actually read the whole thing.

Built a full retrieval-augmented generation pipeline: PDF parsing → vector embeddings → semantic search → LLM-generated answers with page-level citations. Session history and PDF export make it a complete research workflow.

React.jsNode.jsGemini APILangChainTailwind CSS

live source

featured

Crontab

2026

A cloud-native cron job scheduler with AI-powered natural language scheduling. Tell it 'every Monday at 9am' and it figures out the cron expression, manages HTTP pings, and logs every execution.

Needed a way to schedule recurring HTTP health checks without wrestling with cron syntax. Built a full scheduling platform with JWT auth, Argon2 password hashing, per-user job isolation, and an AI landing page.

Multi-tenant • Secure by default • AI-enhanced UX

TypeScriptBunReact.jsExpress.jsShadcn

live source

Private Chat

2025

A secure, real-time messaging platform built on Elysia.js and Bun for raw speed. Messages flow through Upstash Redis — no data touches a traditional database, so nothing persists unless you want it to.

Explored the Bun + Elysia.js runtime for a latency-obsessed use case. WebRTC handles peer connections, Redis Pub/Sub manages message routing, and the whole thing is deployed on edge infrastructure.

Zero-persistence by design • Edge-deployed • Sub-100ms delivery

Next.js 16Elysia.jsBunUpstash RedisWebRTCTypeScript

live source

featured

Donkey Type

2026

A gorgeous, minimalist typing test app obsessed with 60fps animations and fluid UX. Every keystroke feels buttery because every state transition is memoized and component-driven.

Wanted to prove that a React app could feel as smooth as a native desktop tool. Zustand keeps state predictable, Framer Motion handles micro-animations, and every re-render is surgically scoped.

2 ⭐ on GitHub • 1 fork • Designed for speed, engineered for feel

React 19ZustandTailwind v4Framer MotionTypeScript

live source

Drop It

2026

Ultra-fast, zero-server P2P file sharing. Files travel directly between browsers via WebRTC data channels — no server ever touches your data. Just open, share a link, and drop.

Built to solve the 'how do I send this file to someone across the room' problem without cloud uploads. WebRTC handles the heavy lifting, a signaling server brokers the initial handshake, then gets out of the way.

Peer-to-peer • No file size limits • Privacy-first

Next.jsWebRTCTypeScriptTailwind CSS

live source

IPL Qualification Analyzer

2026

A Telegram bot that answers cricket fans' burning question: 'Can my team still qualify?' Uses an LLM with live data tools to calculate points, NRR, and explain exactly what each team needs.

Combined Telegram Bot API with function-calling LLMs to create an AI agent that reasons about live cricket standings. The bot fetches real-time data, runs qualification simulations, and explains results in natural language.

AI Agent • Live data • Function calling

TypeScriptTelegram APILLMFunction Calling

live source

Video Transcoding Service

2026

A backend microservice for video transcoding — upload a video, queue it for processing across multiple resolutions, and stream the result. Built to understand media pipelines at a systems level.

Dug into ffmpeg, job queues, and chunked uploads to build a service that handles video format conversion at scale. The focus was on understanding how platforms like YouTube handle the upload → transcode → serve pipeline.

Systems engineering • Queue-based architecture

TypeScriptNode.jsFFmpegJob Queues

source