projects
Multimodal AI: Image Caption Generation System
- Developed a scalable CNN-LSTM deep learning system in Python to automatically generate contextual text descriptions for images.
End-to-End Speech Recognition Web Service
- Deployed an end-to-end deep learning speech recognition system with an interactive Streamlit web interface for real-time transcription.
HealthTech AI: Medical Image Segmentation
- Engineered a containerized U-Net medical imaging application with a FastAPI backend and intuitive UI for semantic segmentation of breast cancer pathology.
Deep Learning for Breast Cancer Pathology Image Segmentation
- Implemented and evaluated a U-Net-based deep convolutional architecture for precise semantic segmentation of breast cancer pathology images under the supervision of Dr. Mahdi Jabalameli.
Reverse Engineering: Video Borescope Device
- Co-designed and reverse-engineered industrial video borescopes featuring interchangeable probes (2.4mm--8.4mm) and 4-way joystick articulation for advanced machinery endoscopy.