PhishNet: Phishing Website Detection using Federated Learning (project)
Contributed to building a deep neural network–based phishing URL classifier for privacy-focused detection using federated learning concepts. Focused on preparing and using labeled inputs to train/evaluate the model for classifying phishing versus legitimate URLs. Worked on end-to-end experimentation in a cybersecurity ML workflow to support model performance improvements. • Labeled/organized training data representations for URL-based classification. • Implemented and iterated model training and inference logic in Python. • Tested hypotheses to improve classification quality for phishing detection. • Integrated components in a federated-learning-style pipeline setup.