ALI SAMET

Developed an AI-powered mobile application for real-time body movement detection using the smartphone camera. Integrated Google ML Kit’s human pose detection model. Implemented secure authentication using Firebase Auth. Designed and developed an automatic algorithm to count physical exercise repetitions (squats, push-ups, sit-ups) in real time. Conducted a comparative study of state-of-the-art human pose detection models (MediaPipe Pose, OpenPose, Google ML Kit Pose). Designed and built a responsive and interactive user interface with Flutter and custom widgets. Implemented CRUD operations for fitness exercises (create, read, update, delete) using Cloud Firestore. Applied Bloc/Cubit architecture for state management and modular code organization.

-Developed and customized a desktop application for an existing marketplace website dedicated on on-demand manufacturing, resulting in a 30% increase in user utilization speed. -Managed user access to various pages of the application, this streamlined user navigation and ensured secure access, contributing to a more organized and efficient user experience. -Built a CI/CD pipeline, reducing deployment time by 40%. -Deployed the desktop application to Microsoft Azure. -Added a new feature to the website, enabling users to download the desktop application for different operating systems (Mac, Windows, Linux). -Added new features to the admin portal, including interactive charts for data visualization.

-Collected over 100,000 samples of data for different use cases (email classification, burnout detection, sentiment classification and detection of abusive language in images). -Conducted data cleaning, visualization, and analysis. -Compared multiple Machine Learning and Deep Learning algorithms to evaluate their performance on different use cases. -Enhanced models prediction and classification accuracy by 20% by implementing new data preprocessing techniques. -Deployed suitable models as APIs (Web Services) using FastAPI and Docker. -Implementated a text language detection algorithm.