• 🥇 Winner, AIML Hackathon under Culrav/Avishkar at MNNIT Allahabad 2024
• 🥇 Winner, Hactivate, under Botrush 3.0 conducted by Robotics Club MNNIT ALLAHABAD 2025
• 🥇 Winner, Galactic Heist, under Botrush 3.0 conducted by Astronomy Club MNNIT ALLAHABAD 2025
• 🥇 Winner, Robowars, under Botrush 3.0 conducted by Robotics Club MNNIT ALLAHABAD 2025
• 🥈 Runner-up, Quinthalon - Mock Interview Process conducted by ES Society MNNIT ALLAHABAD 2024
• 🏆 Second Runner-up, DEVJAM under Weekend of code, MNNIT Allahabad 2025

CampusBeacon aims to be the central hub for campus life, connecting students, faculty, and alumni seamlessly.

• Core Features:
  • 📢 Real-time Announcements & Event Feeds: Stay updated with live notifications for campus news, events, workshops, and deadlines.
  • 📚 Resource Sharing Hub: Centralized access to notes, assignments, past papers, and project repositories shared by peers and faculty.
  • 👥 Club & Community Management: Dedicated spaces for clubs and societies to manage members, post updates, and organize events.
  • 💬 Discussion Forums & Q&A: Engage in academic or social discussions, ask questions, and share knowledge within the campus community.
  • Lost & Found, Ride Sharing, Buy And Sell and a lots more to come!
• Goal: To enhance collaboration, communication, and resource accessibility within the MNNIT Allahabad ecosystem (and potentially other institutions).
• Tech Stack: React, Node.js, Express.js, MongoDB, PostgreSQL, Supabase (for Auth/Realtime features), TailwindCSS, REST APIs.
• Links: GitHub Repository | Live Website

DAIOS is a distributed, agent-based orchestration platform for autonomous software development, DevOps, and intelligent workflow automation.

• Core Features:
  • 🧠 LLM-Powered Agents: Specialized agents for development, debugging, operations, and orchestration—each leveraging large language models for reasoning and task execution.
  • 🛠️ Pluggable Tooling: Extensible toolbox for code generation, file ops, semantic search, CI/CD, shell, networking, and more.
  • 🔁 Task Decomposition & Planning: Breaks down complex tasks into actionable subtasks using advanced planning skills and LLMs.
  • 🔗 Inter-Agent Communication: Robust message bus (Redis pub/sub) for decoupled, scalable agent interactions.
  • 📦 Integrated Vector Search: Chroma vector DB for context retrieval, code/document search, and memory.
  • 👀 Observability & Monitoring: Built-in logging and monitoring hooks for agent activity and orchestration flows.
  • ⚡ GPU-Accelerated LLM Inference: Ollama for local, high-performance model serving.
  • 🔒 Auth, API & Web UI: Secure REST API, extensible with web dashboard and OAuth support (upcoming).
  • 🚀 Composable, Modular, Cloud-Native: Designed for scale, extensibility, and seamless integration into modern workflows.
• Goal: To empower teams and individuals to automate, accelerate, and intelligently orchestrate complex software and operations workflows—end-to-end.
• Tech Stack: Node.js, Python, Redis, Docker, FastAPI, Express.js, React, TypeScript, Chroma, Ollama.
• Links: GitHub Repository | Live Dashboard (Upcoming)

An intelligent, full-stack environmental monitoring system built for Hack36 under the mentorship of the Computer Coding Club, MNNIT. Project ULTRON empowers users to track ecological changes in real time and receive early warnings about potential environmental threats.

• Key Features & Design:

  • 🗺️ Custom Area Selection: Users can log in and draw any custom-shaped region on the map to monitor.
  • 📬 Environmental Alerts: Subscribe to receive notifications for:
    • 🌳 Deforestation
    • ❄️ Glacier Melting
    • 🔥 Fires
    • 🌊 Coastal Erosion
    • 🌧️ Flooding
  • 🛰️ Google Earth Engine Integration: Real-time satellite data analysis with indices like NDVI, NDWI, etc.
  • ⚡ High-Performance Backend: Direct execution of Python scripts from the JavaScript backend via child_process, avoiding Flask HTTP overhead.
  • 📚 Resource Center: Guides users with best practices during ecological emergencies.
  • 📈 History Page: Explore how selected regions have changed over time — perfect for NGOs, researchers, and environmentalists.
  • 🚀 Space Trackers: Live position trackers for the International Space Station (ISS) and Hubble Space Telescope.
  • 🛠️ Analysis Dashboard: Full CRUD functionality for managing subscriptions, alerts, and historical insights.

• Tech Stack: Node.js, Express.js, React.js, Google Earth Engine (Python API), Tailwind CSS

• Links:

  • 🌐 Live Website
  • 🎥 Demo Video
  • 🛠️ GitHub Repository

A powerful, flexible profanity filter for modern JavaScript and TypeScript applications, supporting multiple languages and scripts out-of-the-box.

• Key Features & Design:

  • 🌎 Multi-Language Support: Detects profanity in English, Hindi, Hinglish, Bengali, Tamil, Telugu, French, German, and Spanish.
  • 🧠 Multiple Scripts: Works seamlessly with Latin/Roman and native scripts (Devanagari, Bengali, Tamil, Telugu).
  • 🧹 Flexible Cleaning Options: Supports character-level or word-level placeholder replacement for detected profanities.
  • ✏️ Customizable Filters: Dynamically add, remove, and modify words, and set custom placeholders.
  • ⚡ High Performance: Optimized for large-scale content filtering with minimal overhead.
  • 🛡️ Zero Dependencies: Only uses leo-profanity as a base, ensuring a lightweight package.
  • 🛠️ TypeScript Native: Built with TypeScript, providing full type safety and IntelliSense support.
  • 🔮 Extensible Architecture: Easily extend to new languages or custom detection methods.

• Tech Stack: TypeScript, Node.js, Regular Expressions

• Links: NPM Package | GitHub Repository

An interactive, deep learning-based sentiment analysis system inspired by the "Money Heist" series, built for BotRush 3.0 to detect emotional states from voice recordings in real-time.

• Key Features & Design:

  • 🎙️ Surveillance Records: Upload and analyze pre-recorded audio files (.wav, .mpeg, .amr) for emotion detection.
  • ⚡ Quick Interrogation: Record 7-second voice samples on the fly for immediate emotion classification.
  • 🛰️ Live Monitoring: Continuously capture and process microphone audio to track emotional trends in real-time.
  • 🧠 Deep Learning Models: Custom CNN trained on MFCC and mel-spectrogram features for high-accuracy emotion recognition.
  • 🔄 Threaded Audio Processing: Implements threading for smooth, responsive real-time monitoring without UI lag.
  • 📊 Dynamic Visualizations: Real-time updating of waveforms, MFCCs, mel-spectrograms, and emotional trend graphs.
  • 🚨 Alert System: Instantly flags distress or unstable emotional states during live monitoring sessions.
  • 🎨 Themed UI: Built with Streamlit, featuring a custom Money Heist-inspired visual design and user flow.

• Tech Stack: Python, TensorFlow, Streamlit, Librosa, SoundDevice

• Links: GitHub Repository

A modern URL shortener service demonstrating advanced system design principles for scalability and performance.

• Key Features & Design:
  • 🛡️ Quantum-Resistant Hashing: Uses SHA-3 (Keccak) for secure and unique short URL generation.
  • ⚡ High-Performance Caching: Multi-level caching (Redis + In-Memory) ensures ultra-fast URL redirection.
  • 📈 Scalable Architecture: Designed with principles like load balancing and read replicas (conceptual) for high throughput.
  • 🚦 Reliability Measures: Implements Rate Limiting & Request Throttling to prevent abuse, plus Circuit Breaker patterns for fault tolerance.
  • 📊 Analytics Dashboard: Tracks click-through rates and provides basic analytics on URL usage.
• Tech Stack: Node.js, TypeScript, MongoDB, Redis, Express.js.
• Links: GitHub Repository | Live Website

A modern, high-performance URL shortener package with quantum-inspired encryption for enhanced security and privacy protection.

• Key Features & Design:

  • 🔐 Multi-Layer Security: Implements quantum-inspired noise generation with triple-layer AES encryption for unmatched URL protection.
  • ⚙️ Highly Configurable: Customizable URL length (4-100 chars), salt rounds, and quantum seed parameters for tailored security levels.
  • 🔄 Flexible Storage: Works with any database system including PostgreSQL, MongoDB, and Redis with full TypeScript support.
  • ⚡ Performance Optimized: Supports concurrent operations with minimal variance and excellent scalability for high-traffic scenarios.
  • 🧪 Production Ready: Comprehensive test coverage (98%+) with robust error handling and sanitization for reliable implementation.

• Tech Stack: TypeScript, Node.js, AES Encryption, PBKDF2, SHA3

• Links: NPM Package | GitHub Repository

A modern, scalable implementation of Tic-Tac-Toe in C++ showcasing advanced algorithms, flexible configurations, and smart AI mechanics.

• Key Features & Design:

  • 🔄 Dynamic Game Engine: Supports custom board sizes (n x n) and win conditions (k-in-a-row) for endless variations.
  • 🧠 AI With Depth: Three difficulty levels using Random logic, Minimax, and Alpha-Beta Pruning. Includes an opening book for optimized 3x3 play.
  • 👥 Multiple Game Modes: Play solo vs AI or challenge a friend in multiplayer mode.
  • 💾 Stateful Gameplay: Save/load game sessions, undo last moves, and track scores across rounds.
  • 🖥️ Responsive CLI UI: Auto-scaling terminal board display with input validation and user-friendly command system.

• Tech Stack: C++ (STL,File I/O), Minimax Algorithm, Alpha-Beta Pruning

• Links: GitHub Repository

A sophisticated chess implementation featuring real-time machine learning-based position analysis and a professional graphical interface.

• Key Features & Design:

  • 🎮 Interactive Gameplay: Fully functional chess engine with legal move validation, pawn promotion, and intuitive controls.
  • 🧠 AI Position Analysis: Real-time board evaluation using RandomForest machine learning model trained on professional games.
  • 📊 Visual Advantage Bar: Dynamic visualization of position strength with gradient-based advantage representation.
  • 🎨 Professional UI: Clean, modern interface with responsive design, hover effects, and visual feedback mechanisms.
  • 🧩 Modular Architecture: Well-structured codebase with separate modules for game logic, AI engine, and interface components.

• Tech Stack: Python, Pygame, Python-Chess, Scikit-Learn, NumPy

• Links: GitHub Repository | Documentation

A robust library management system built in C++ featuring efficient book tracking, borrower management, and persistent data storage with optimized lookup algorithms.

• Key Features & Design:

  • 🏛️ Complete Library Workflows: Integrated borrowing, returning, and cataloging operations with intuitive command-line interface.
  • 🔍 Optimized Data Retrieval: Implements hash table (unordered map) for O(1) constant-time book lookups by ID and title.
  • 🔐 Administrative Controls: Secure admin panel with password protection for inventory management and system configurations.
  • 💾 Persistent Storage: File I/O operations for data preservation between sessions with pipe-delimited format for efficient parsing.
  • 🧩 Modular Architecture: Object-oriented design with specialized classes for books, records, and student management.

• Tech Stack: C++, STL Containers, File I/O, Hash-Based Data Structures

• Links: GitHub Repository | Documentation

A robust backend service powering interactive chess applications with real-time gameplay, move validation, and seamless player communication.

• Key Features & Design:

  • 🔄 Real-time Communication: Socket.IO implementation enabling instant move updates and game state synchronization between players.
  • ⚙️ Professional Game Logic: Leverages chess.js for comprehensive rule enforcement, position validation, and game termination detection.
  • 🖥️ Server-Side Rendering: EJS templating for dynamic view generation, creating a cohesive integration between game logic and UI.
  • 📱 API Architecture: RESTful design principles with Express for game management operations and player interactions.
  • 🛠️ Developer-Optimized: Configured with nodemon for rapid development cycles and efficient debugging processes.

• Tech Stack: Node.js, Express, Socket.IO, chess.js, EJS

• Links: GitHub Repository | Documentation

A computer vision-powered robotic solution that identifies ground patterns and automatically traces them using precise marker control, built for MNNIT's Avishkar Tech Fest.

• Key Features & Design:

  • 👁️ Real-time Pattern Recognition: Advanced OpenCV implementation for instant pattern detection from live camera feeds or static images.
  • 🖊️ Dynamic Marker Control: Precision drawing mechanism with automated motor control for accurate pattern tracing.
  • 🧠 ArUco-Based Positioning: Computer vision algorithms for marker identification and orientation tracking with spatial awareness.
  • ⚡ Hardware Integration: Seamless communication between Python processing system and Arduino-controlled motors via serial interface.
  • 🔄 Adaptable Performance: Customizable dictionary-based pattern recognition with adjustable parameters for various environmental conditions.

• Tech Stack: Python, OpenCV, NumPy, Arduino, Serial Communication

• Links: GitHub Repository | Documentation

A performance-optimized media gallery showcasing Holi celebration memories with rich interactive features, smooth animations, and responsive design.

• Key Features & Design:

  • 🖼️ Responsive Media Grid: Dynamically adjusting layout displaying images and auto-playing video thumbnails for optimal viewing.
  • ❤️ Interactive Engagement: Like and reaction system with colorful confetti animations for an engaging user experience.
  • 🔍 Enhanced Media Viewing: Modal-based media playback with fullscreen toggle and intuitive navigation controls.
  • 📱 Modern Sharing Options: Integrated Web Share API for native sharing capabilities across devices and platforms.
  • ⚡ Performance Optimized: Memoized components and optimized callbacks ensuring smooth interactions even with heavy media content.

• Tech Stack: React, Framer Motion, Canvas Confetti, React Icons

• Links: GitHub Repository | Live Website | Documentation