Hey, I'm Aditya 👋

I work as a Software Developer at
Konrad LogoKonrad
, focusing on full-stack web applications and AI integrations.
I studied Computer Science and Engineering at
JKLU, Jaipur LogoJKLU, Jaipur
, with visiting semesters at
IIT Gandhinagar LogoIIT Gandhinagar
 and
IIT Jammu LogoIIT Jammu
.
I like understanding how things work beneath the surface. I focus on clear abstractions, practical design, and building systems that behave predictably.
Outside of work, I follow commercial aviation and photograph environments that show structure or symmetry. I also explore how real-world systems work, including economics and geopolitics, mostly out of curiosity rather than depth.
what I'm doing now
Latest Blog Post

Why I built a web search library

I was working on a side project where I needed to add basic web search capability to an AI agent. When I looked around, I noticed most projects rely on existing search APIs like Bing, Google, Perplexity, Brave, etc. For my use case, that felt like ov...

3 min read
Dec 29, 2025
Read more

Projects - Open Sourced

Some of my recent projects.

Quack Search

A DuckDuckGo web search library for JS runtimes, written in Go.

Go
Library
Web-search
GitHubLive Demo
Gollama

MCP powered chat interface to track issues, update code, and open PRs. 

Go
React
Agents
MCP
GPT-OSS
GitHub
Relay - Super Efficient URL Shortener

URL shortener with caching, rate limiting, and serverless deployment. 

Go
Gin
PostgreSQL
Redis
GitHubLive Demo
Chessify - Online Chess

Chess platform supporting play against Stockfish AI or next available user.

Express
React
WebSockets
Stockfish
AWS
GitHubLive Demo
SafeXchange

File sharing with email verification, roles, and encrypted links. 

FastAPI
Next.js
AWS S3
MongoDB
Docker
GitHub
removeSillyBG - Instant BG Remover

Perfect for quickly removing background from images.

FastAPI
Docker
Python
GitHub
VisBubble - P2P Video Calling Platform

P2P video calling app using WebRTC and real-time signaling with Socket.IO. 

Express
React
WebSockets
WebRTC
GitHub
EmoVision

Real-time emotion recognition, 4 layer CNN for emotion recognition achieved 73% precision on a 35K image dataset.

TensorFlow
Scikit-learn
JupyterLab
GitHub
See more on GitHub