aya saadawi

- Developed a hybrid computer vision system combining keypoint detection for analogue gauges and object detection for digital devices to accurately read industrial instrument data. - Built a full end-to-end AI workflow, including data collection, preprocessing, model training, evaluation, and integration into a real-time deployment pipeline.

- Developed an end-to-end AI transcription system for French radiology reports that surpassed OpenAI Whisper STT accuracy, including data collection, preprocessing, fine-tuning, backend API, and frontend interface. - Automated the entire workflow using n8n, integrating multiple APIs and services for seamless data ingestion, transcription, and report generation.

- Developed a credit scoring model to predict individuals’ creditworthiness using classification algorithms. - Optimized model performance and accuracy through feature engineering, hyperparameter tuning, and evaluation across multiple ML algorithms.

- Developed a full-stack basic banking system, enabling users to view and transfer money between customers. - Deployed the application on Heroku, using Docker for the frontend and Heroku-specific SQL for database management. - Refined skills in UI/UX design, debugging, and application testing.