I build full-stack applications, real-time systems, and AI-native products - from agentic app builders to AI presentation tools.
Read my writingA little context
I graduated from IIT (ISM) Dhanbad with a B.Tech in Minerals and Metallurgical Engineering. My path into software wasn't a conventional CS one - I developed my skills by building, breaking, and figuring things out.
That led to full-stack applications, serverless systems, real-time collaboration, and agentic products. Along the way, I worked remotely as a JavaScript / TypeScript expert in the Scale AI ecosystem, contributing to LLM evaluation and feedback workflows.
Now I'm going deeper into AI application engineering, modern AI, and machine learning - without losing the habit of shipping things.
How I think about building
Select an intersection ↴Connected by curiosity. Grounded in engineering.
Turning model capabilities into usable products, while exploring the architectures underneath.
LLMs · Agents · Structured generation · ML
Work & perspectives
Remote
Expert human feedback and programming data for LLM training and evaluation workflows.
Saint Petersburg, Russia
Selected for a UNESCO-backed international summer school on management of the mineral resource sector, including industrial visits and applied case-study work.
Built, not just imagined
A collection of systems, products, and experiments I've built.
An agentic prompt-to-app platform that turns natural-language ideas into working React applications.
Converts natural-language prompts into structured presentations, generates visual assets, and provides an editable slide workspace with PowerPoint and PDF export.
A real-time collaborative drawing environment built around WebSockets and a Turborepo monorepo architecture.
An AI-powered mock interview platform supporting generated questions, voice interviews, transcription, and structured interview sessions.
A space to play
Things I build when I don't have a good reason to.
Recursive geometric patterns rendered entirely with HTML Canvas.
Experiment coming soon ↗Interactive 3D scenes and WebGL experiments.
Experiment coming soon ↗Small experiments while learning modern machine learning and deep learning.
Experiment coming soon ↗The engineering notebook
Notes from building and learning.
Deploying a web app to an AWS EC2 machine with Nginx and Certbot, including running multiple apps on one server.
Sep 25, 2025How wallets derive public-private key pairs, the role of ECDSA and EdDSA, and what HD wallets are.
Jun 29, 2025A hands-on guide for React developers moving to Next.js: App Router, API routes, server actions, and file-based routing.
Jun 18, 2025Always a work in progress
Attention · Transformers · LLM architectures
Deep Learning · CNNs · Representation Learning
Agents · Tool Use · Structured Generation · Streaming
Three.js · WebGL · Generative Art
Along the way
Inter-college CTF Competition
IIT (ISM) Dhanbad Annual Hackathon
Max Rating: 1365
Max Rating: 1590
Where it started
B.Tech - Minerals and Metallurgical Engineering
CGPA: 8.71 / 10
A different starting point. The same curiosity.Beyond Code
Sometimes I also over-engineer things that absolutely did not need engineering.