NexFlow AI: An Urban Nervous System for traffic. Using Graph Neural Networks (GNN) and Graph Attention (GAT), NexFlow solves hyper-dense Indian roadway congestion. It enables real-time junction communication and absolute emergency vehicle priority. Built with PyTorch Geometric and SUMO, it’s a production-ready, OpenEnv-compliant RL controller.
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Learn it. Build it. Ship it for others.
An advanced AI Engineering Copilot built on a distributed architecture. It leverages a Node.js orchestrator and Redis-backed job queue to manage high-volume repository indexing. The Python AI engine uses AST-level chunking and multi-modal vision for deep codebase intelligence. Built for scale, security, and 'self-healing' infrastructure logic.
Apache Airflow - A platform to programmatically author, schedule, and monitor workflows
AI coding assistant skill (Claude Code, Codex, OpenCode, Cursor, Gemini CLI, and more). Turn any folder of code, SQL schemas, R scripts, shell scripts, docs, papers, images, or videos into a queryable knowledge graph. App code + database schema + infrastructure in one graph.
Python codes
This project is a decentralized Lottery system built using Solidity and the Hardhat development framework. It allows users to enter a lottery by paying a fixed entry fee of 0.01 ETH, and a manager can draw a random winner who receives 95% of the prize pool. The remaining 5% is retained as a platform fee.
ParlimentAI decodes Indian Parliament for 1.4B citizens. Browse bills, watch live debates, ask questions in 22 languages, and get AI-powered plain-language summaries showing exactly how every new law affects your life — farmer, soldier, student, or business owner.
NexFlow-GNN is an AI-powered traffic signal controller tailored for complex Indian city intersections. Built with Reinforcement Learning and PyTorch, it acts as an "Urban Nervous System" to dynamically manage traffic flow. It adapts to rush hours, prioritizes emergency vehicles, and reduces overall congestion via the OpenEnv simulation.
Jarvis is a production-level AI OS layer designed to surpass agents like OpenClaw through multimodal autonomy. Built with Python, it features a Syntax-Aware RAG system for deep codebase intelligence, real-time screen perception via OCR, and verified system automation. It bridges the gap between vision and action to provide proactive debugging.
A horizontally scalable distributed job queue built with Node.js, BullMQ, and PostgreSQL. Features an integrated AI diagnostics engine using Gemini 2.5-Flash to automatically analyze and summarize complex task failures. Supports priority queuing, real-time progress tracking, and Docker-ready orchestration for production environments.