ehab assaf

• developed an xgboost model for keyword identification by leveraging data extracted from sql databases and amazon s3 for advanced feature engineering. containerized the model using docker for seamless deployment and continuous optimization, driving a 125% boost in sales through targeted keyword strategies.
• engineered a full-stack data extraction solution by integrating beautifulsoup, selenium, and data scraping apis with a flask-based web interface (html/css) and firebase backend, reducing manual data collection by 65% while enabling real-time dashboard updates.
• developed a microservices-based full-stack platform for automated amazon listing creation, leveraging a flask back-end, react front-end, and docker/kubernetes orchestration. integrated asynchronous data ingestion from sql and amazon s3 with ci/cd pipelines via github actions, reducing listing creation time by 400%.

• developed interactive tableau dashboards by leveraging advanced sql queries to extract, clean, and transform amazon product performance data—reducing reporting turnaround time by 40% and empowering management with actionable insights.
• led a/b testing initiatives, optimizing product bundling strategies that enhanced sales performance and improved product-market fit, while managing resource allocation and monitoring key performance indicators.
• developed automation tools leveraging sql and machine learning to optimize reporting processes, improving data accuracy and increasing efficiency by 25%.

• refactored, documented, and debugged python-based elt data preprocessing pipelines handling sensitive hospital patient data, employing git version control for seamless code integration.
• performed unit tests on preprocessing pipelines, enhancing data management, logic, and readability; resulting in improved system efficiency across 50 infusion centers with a total of 775 infusion chairs.
• built dynamic dashboards using tableau to measure kpis across 50 infusion centers, improving performance tracking and decision-making.

aug 2021 - may 2024

may 2024 - dec 2025

• led the development of a machine learning model to predict song popularity on spotify, utilizing libraries like scikit-learn, tensorflow, and pytorch to apply advanced feature engineering and supervised/unsupervised learning techniques.
• designed dashboards using superset to visualize song popularity trends, providing actionable insights for content strategy optimization.
• collaborated cross-functionally with data scientists, ensuring statistical rigor through model validation and optimization with scikit-learn, delivering insights for future hit song identification and improving recommendation algorithms.

• developed and deployed a logistic regression classifier in python using sklearn, achieving 98% accuracy and reducing spam emails by 75% through advanced feature engineering.
• employed robust cross-validation and regularization techniques to mitigate overfitting, while integrating explainable ai approaches to deliver actionable insights.
• collaborated with cross-functional teams to transform complex content classification challenges into tractable analytical problems, aligning model metrics with business objectives.

• developed a highly accurate decisiontreeclassifier model to predict titanic survival outcomes, achieving an impressive 91% accuracy by utilizing advanced feature engineering, hyperparameter tuning, and focusing on critical variables like age, gender, and class.
• conducted comprehensive exploratory data analysis (eda) using pandas, numpy, and seaborn, identifying key survival patterns and engineering new features to boost model performance and explainability.