Projects
Fido Based Password Less Authentication For E-Bikes
- Designed and implemented a secure passwordless authentication system using FIDO2/WebAuthn, eliminating traditional passwords and reducing phishing attack surfaces.
- Built a secure web application implementing passwordless login flows, integrating frontend JavaScript with backend APIs.
WebAuth
JavaScript
HTML
CSS
Node.js
express.js
MongoDB
Phishing Website Detection Using a Random Forest Model
- Engineered a Random Forest–based phishing detection model using 23+ handcrafted URL and domain features and built a custom feature extraction pipeline to analyze lexical, structural, and domain-level attributes of URLs and Improved detection reliability through data preprocessing, feature selection, and model evaluation metrics and deployed the model as a Flask web application on Render, enabling real-time phishing classification.
Python
Random Forest model
Flask
TensorFlow/Keras
NumPy
Pandas
Matplotlib
Google Colab
Phishing Url Blocker
- Developed a real-time Chrome browser extension that blocks malicious URLs before page load, enhancing user security and Integrated a Flask-based ML backend with the extension to perform live phishing predictions.
JavaScript
HTML
CSS
Chrome Extension APIs
Random Forest
Flask (backend)
Smart Road Accident Detection & Severity Analysis System
- Developed an intelligent traffic safety system using YOLO for real-time vehicle/accident detection from CCTV or dashcam video streams.
- Integrated a CNN model to classify accident severity levels (minor, moderate, severe) for faster emergency response and decision-making.
- Improved monitoring efficiency by combining fast object detection with accurate severity analysis, enabling automated road incident alerts.
Python
YOLO
CNN
OpenCV
TensorFlow/Keras
NumPy
Pandas
Matplotlib
Google Colab
Gradio