- Languages: Python, C++
- Robotics: ROS2, OpenCV
- AI/ML: PyTorch, TensorFlow
- DevOps: Docker, Git, CI/CD
- Others: Ubuntu, Raspberry Pi, Arduino
- Robot Vision Feedback Loop - Industrial Anomaly Detection - Unsupervised anomaly detection system for industrial inspection using PatchCore + ResNet50. Learns normal patterns only and detects defects as deviations, making it suitable for real-world factory environments with unknown or rare failures.
- Image Segmentation and Representation Learning - image segmentation and feature extraction pipelines, including autoencoder based representation learning to support downstream tasks such as object recognition and scene understanding.