May 2025 – Jul 2025 · 2 mos

• Developed a hybrid system that combines PHQ-8 questionnaire responses with audio speech features to detect early signs of depression. The approach integrates Quantum Variational Classifier (QVQC) and Quantum Convolutional Neural Networks (QCNN) to visualize and analyze quantum decision boundaries, enhancing model interpretability and performance. Audio data was processed using Whisper and embeddings generated with Sentence Transformers, while classical data handling leveraged Pandas and Sklearn. The system is deployed through Streamlit, providing an interactive interface for users to input PHQ-8 scores and speech samples, generating real-time predictions for early intervention.