Hassna EL-BOUSIYDY

● Applied statistical and machine-learning methods to large, complex corpora of unstructured scientific literature, structuring raw text into machine-readable entities, relations and properties for downstream analysis. ● Designed iterative experimental workflows - annotation, training, error analysis - evaluating results with precision, recall and F1 to validate model quality and guide refinement

Conducted research and engineering in machine learning and NLP as part of a PhD candidacy.

- Engineered and managed complex structured and unstructured data pipelines, with a focus on scalability and reproducibility
- Collaborated on early-stage research applying CNNs and image processing techniques for multimodal data fusion
- Applied learning-based parameter tuning to optimize preprocessing pipelines for large-scale text and image datasets, improving feature quality by 60%
- Developed

Designed end-to-end data pipelines in Python and SQL, applying ETL/ELT logic with automated testing. Used Git and Docker for version-controlled workflows.

- Developed neural network models in Python to forecast wave force dynamics, increasing prediction accuracy by 95% and reducing error rates in marine forecasting systems
- Designed and evaluated Transformer-based, LSTM and CNN architectures for long-horizon time series forecasting, enabling flexible short- and long-term trend analysis
- Built modular ML pipelines using Python, PyTorch, and Pandas

Analysed sensor data and benchmarked five deep-learning architectures, achieving 95%+ improvement in predictive accuracy. Built reproducible pipelines and delivered technical documentation to non-specialist teams.

- Built Monte Carlo simulation models to optimize key parameters in lithium-oxygen battery research, improving evaluation accuracy by 60%
- Automated simulation workflows using Python, boosting computational efficiency by 70% and accelerating large-scale data analysis
- Designed iterative learning-based frameworks for parameter tuning, enabling robust convergence to optimal configurations
- Conduc

Developed numerical models of lithium peroxide growth in Li-air batteries using Python and COMSOL, improving simulation accuracy by 30%.

- Cleaned and modeled a massive national census dataset (millions of rows), using R, SQL and SPSS to detect anomalies and enhance data integrity
- Built regression and time series models to identify demographic trends, improving policy recommendation accuracy by 80%
- Automated generation of 25+ key demographic indicators (e.g., age pyramids, median age, dependency ratios), supporting high-impact

Applied statistical methods to Morocco's national census dataset, automating production of 25+ demographic indicators. Built interactive dashboards using R, SPSS, and SAS.

This Master’s program provides advanced training in applied mathematics, focusing on mathematical modeling, numerical analysis, and stochastic processes. Developed in collaboration with the CNRS Laboratory (LAMFA – UMR 7352), the program emphasizes solving real-world problems in areas such as scientific computing, energy storage, ecology, and data analysis. Core competencies include PDE analysis,

Engineered a comprehensive, multidisciplinary program in statistics, applied economics, and data science at the prestigious National Institute of Statistics and Applied Economics (INSEA) in Rabat, one of Morocco’s leading grandes écoles under the High Commissioner for Planning. The curriculum included advanced coursework in statistical modeling and econometrics, survey methodology, data mining, ma

Master's program in applied mathematics focusing on mathematical modeling, numerical analysis, and stochastic processes.

Master of Science in Applied Mathematics.

Doctor of Philosophy in Machine Learning.