Engineer. A CS Nerd who enjoys building and exploring DevOps, Backend, AI, and ML, among other intriguing technologies.
- open-rabbit - Open-source code rabbit implementation (building). A modern AI-powered coding assistant.
- Appwit - Open source alternative to Lovable. Build beautiful applications with an intuitive interface.
- CruxML - A minified Machine Learning and Deep Learning Framework/Library built from scratch in Python.
- Attention-Is-All-You-Need - Transformer architecture implementation. Transformers are the new blood for Neural Networks.
- Speech-emotion-detection - AI-powered system that detects emotions from speech using deep learning with a beautiful web interface for real-time emotion recognition.
- scalable-rag - Scalable Retrieval-Augmented Generation system implementation.
- OpenFoodFacts-MCP - MCP server for managing, processing, and analyzing food product data using Open Food Facts APIs.
- nEn - Graphical AI workflow runner inspired by n8n. Build and visualize complex AI workflows with ease.
- exchange - Full-stack cryptocurrency exchange platform (CFD) with Docker containerization. Features modern TypeScript architecture with separate backend and frontend services.
- dhaniverse - A gamified open-world for Gen Z to learn and manage finance, powered by Web3 identity and on-chain rewards.
- SolSentry - AI-powered Solana wallet and transaction analysis agent to detect fraud and transaction patterns.
- path-filter - 📦 Published GitHub Action - Enables conditional execution of workflow steps based on changed files in PRs and commits. Supports monorepo setups with advanced filtering. Available on GitHub Marketplace.
- DebugBuddy - VS Code extension that detects syntax errors, explains issues clearly, and suggests smart fixes on hover.
- fumadocs-transpiler - A powerful Node.js transpiler that converts annotated Markdown files to Fumadocs-compatible React components.
- vimrc - Custom Vim configuration for zsh and macOS.
- cuda-sgemm - Matrix multiplication CUDA Kernel using tiled matrix multiplication algorithm for transformer inference (attention layers, feed-forward layers).



