My Projects

A collection of software projects showcasing full-stack development, machine learning, and systems programming expertise.

Ride Share Platform
June 2025 - August 2025

Ride Share Platform

Built a comprehensive ride-sharing platform using the MERN stack (MongoDB, Express, React, Node.js) with advanced features including JWT authentication, role-based access control, and real-time location tracking. Implemented geohash-based nearest-neighbor matching algorithm for efficient driver-rider pairing and integrated OSRM for optimal route calculation. The platform uses Socket.IO for real-time bidirectional communication between drivers and riders.

Key Highlights

  • 90.4% uptime with containerized deployment
  • Sub-second ride matching latency
  • Real-time location updates
  • Secure authentication with JWT

Technologies

MongoDBExpressReactNode.jsDockerJWTSocket.IOVercel
Cyber Threat Predictor
May 2025 - July 2025

Cyber Threat Predictor

Developed an advanced machine learning system for detecting cyber threats in Industrial IoT networks. Processed and analyzed 2.1 million network packets using optimized preprocessing and feature engineering techniques. Trained an ensemble of SVM and Random Forest models achieving 94.1% detection accuracy with only 0.1s inference latency per sample. Implemented a federated learning pipeline to improve cross-site threat detection while maintaining data privacy.

Key Highlights

  • 94.1% threat detection accuracy
  • Reduced false positives by 34%
  • Processed 2.1M network packets
  • 0.1s inference latency
  • Federated learning for privacy

Technologies

PythonDjangoSVMRandom ForestMachine LearningFederated Learning
ChatterBox
February 2025 - May 2025

ChatterBox

Designed and implemented a robust terminal-based chat application in C++ demonstrating advanced systems programming concepts. Used System V IPC for inter-process communication, implemented RAII wrappers for safe resource management, and employed mutex-protected server-client concurrency to handle multiple simultaneous connections. Built a custom lightweight binary protocol with timestamps ensuring ordered message delivery.

Key Highlights

  • Supports 50+ concurrent users
  • 0.4s average response latency
  • Reduced dropped messages by 93%
  • Custom binary protocol
  • Thread-safe architecture

Technologies

C++System V IPCThreadingRAIIConcurrencyBinary Protocol

Interested in Working Together?

I'm always open to discussing new projects and opportunities.