Skip to main content

Projects

Personal Projects

Chessed Platform

  • Successfully developed a four-player chess platform using React and TypeScript, allowing simultaneous gameplay on a single board
  • Utilized TypeScript interfaces to identify potential errors and React hooks for dynamic component state management and updates
  • Implemented game features including move suggestions, turn validation, piece movement, castling, pawn promotion, and check detection
  • Enabled endgame scenarios with capturing, checkmate detection, ensuring valid moves, game termination when a player is checkmated

GarudaNvim

  • Developed GarudaNvim, a customizable code editor and Neovim distribution written in Lua, enhancing efficiency with custom configurations
  • Implemented features like easy installation with bash scripts, plugin version freezing, Noice for dynamic UI messages and various themes
  • Optimized user experience with over 100 custom keymaps, 25+ plugin configurations, and efficient error handling for development workflows
  • Built a documented site using MkDocs, reflecting installation, features, current and past releases and how-to guides for seamless onboarding

Deribit Trading System

  • Built a high-performance trading system in C++ with multi-threaded WebSocket server for real-time market data streaming and order execution
  • Integrated Deribit Testnet API with OAuth2 authentication, enabling seamless placement, cancellation, and modification of spot, futures, and options orders
  • Implemented performance monitoring tools to capture WebSocket propagation, order processing, and API latency metrics for system optimization
  • Designed a command-line interface supporting live order book visualization, position tracking, and instrument management for efficient trading operations

LinkedinOS

  • Engineered a full-stack Python automation platform using Selenium and BeautifulSoup4 to systematically scrape founder data from Y Combinator and send personalized LinkedIn connection requests.
  • Implemented secure OAuth 2.0 integration with the Google Sheets API to automatically structure and store scraped company and founder data for centralized intelligence.
  • Developed advanced state management to intelligently handle connection request responses (Pending, Already Connected, Email Required) and log outcomes to a JSON file for analytics.
  • Designed a robust, menu-driven CLI architecture with modular components for batch selection, scraping, and connection management, ensuring scalability and ease of use.

SummarizeWebPageAI

  • Engineered a scalable microservices architecture with React, Spring Boot, Scala, and PostgreSQL, ensuring efficient data pipelines and persistent storage for summaries
  • Integrated an AI-driven summarization engine using Python FastAPI and Google Gemini LLM, enabling automated extraction of key insights from unstructured website content
  • Deployed containerized services with Docker and Helm, implementing orchestration strategies aligned with enterprise-grade cloud-native infrastructure standards
  • Delivered user-facing features for real-time summarization, historical search, and data retrieval, supporting intelligent information management and decision-making workflows

Docker Log Streamer

  • Designed and implemented a full-stack web application for real-time streaming, searching, and filtering of Docker container logs using WebSocket and REST APIs
  • Engineered a scalable backend with Node.js, Express.js, and MongoDB Atlas, ensuring persistent log storage and efficient query performance
  • Developed a React.js frontend with dedicated components for live logs, search functionality, and timestamp-based filtering, enhancing user experience
  • Deployed the system on Render and Vercel, integrating containerized log generation with remote streaming for robust and seamless operation

Haunted House

  • Built an immersive 3D haunted environment using Three.js, enabling interactive navigation with realistic lighting, fog effects, and atmospheric rendering
  • Designed detailed models including textured walls, doors, graves, and animated ghost characters, enhancing the overall spooky environment and user engagement
  • Implemented dynamic lighting effects with point lights, spotlights, and animated ghost lights to create eerie and responsive visual experiences
  • Optimized performance through texture mapping, geometry reuse, and efficient rendering techniques, ensuring smooth frame rates across devices

Projects Supervised by a Professor

Cloud-Based Fire Detection and Air Quality System with AI (Under Prof Bhaktha)

  • Designed an IoT-enabled system integrating fire and air quality sensors (MQ2, MQ135) with microcontrollers, enabling accurate hazard detection within 5–10 ft range
  • Developed a cloud architecture using MQTT and MongoDB Atlas for secure, real-time data ingestion, scalable storage, and centralized monitoring across multiple nodes
  • Built a MERN stack dashboard with live logs, data visualization, and downloadable reports, ensuring < 1s latency updates and user-friendly access to historical data
  • Incorporated AI-based anomaly detection with Gemini models and WebSocket-driven alerts, reducing false positives and enhancing overall system reliability

Ayurvedic Consultation Platform (Under Prof Mrigank Sharad)

  • Built a robust backend with NodeJS and ExpressJS, utilizing MongoDB Atlas for efficient database management and data storage
  • Integrated Firebase authentication for secure user access and intelligent categorization based on user type: patient, doctor, or retailer
  • Integrated machine learning algorithms to automatically match patients with doctors based on prakriti determination form inputs
  • Developed a comprehensive payment interface using Razorpay, including API routes for consultation and medicine transactions

Fire and Smoke Detection System (Under Prof Sudip Mishra)

  • Developed an IoT-based fire detection system using ESP8266, MQTT, and MQ2 gas sensors, enhancing smoke detection accuracy
  • Integrated real-time data transmission via MQTT and Mosquitto Broker, ensuring prompt fire alerts through email notifications
  • Utilized machine learning algorithms to reduce false alarms, improving fire identification precision and overall system reliability
  • Addressed market gaps with a dual-layered detection approach, ensuring timely and accurate hazard alerts for enhanced fire safety