Nedal Altiti

Driven Data Science with a passion for transforming complex data into valuable insights that drive business success. Staying up-to-date with the latest tools and techniques in the field, bringing fresh ideas and innovative solutions to the team. Leveraging expertise in statistical methodologies and programming languages, adding value to the company and making a meaningful contribution to the shared goals.

MLOps - Udacity

Orange FTTH Customers Churn Prediction

• Led a team of 3 data scientists in a capstone project to predict customer churn for FTTH orange customers using a real-world dataset
• Performed data visualization using Tableau, data preparation and cleaning, handling imbalance, skewness and outliers, feature selection and engineering.
• Built and evaluated various machine learning models, such as logistic regression, random forest, and XGBoost, and selected the best one based on appropriate metrics.
• Deployed the final model as a simple UI using Docker and FastAPI and documented the code and
  results on GitHub. (https://orange-fiber.netlify.app/)

Image Captioning

• develop a neural network model that can
  automatically generate descriptive captions for
  images.
• train a model using the (MS COCO) dataset and
  evaluate its performance on new images.

Dynamic-Risk-Assessment-System

• The project covers the full scope of ML model deployment with continuous ingestion of data and detection of model drift leading to retraining / redeployment decisions. The model is made accessible through an API with multiple endpoints providing inference and various statistics about the model performance, the data and overall pipeline health.

Justice AI: NER Extraction and Real-Time Speech-to-Text for Legal
Proceedings in Jordan

• This project focuses on developing a comprehensive Natural Language Processing (NLP) solution for the legal justice system in Jordan. It includes the development of a Named Entity Recognition (NER) model to extract key information such as duration, entities, and types of judgments from legal documents. Additionally, it incorporates a real-time speech-to-text model that can be used during legal proceedings to transcribe spoken words accurately and efficiently.