Nadjat Sidi Moussa

- Optimization of AI models and hyperparameters: Adjustment of Random Forest and Decision Tree models to improve forecast performance using techniques like Grid Search, by adjusting parameters such as min_samples_leaf.

- Debugging and validation: Thorough testing, log analysis, and anomaly resolution to ensure reliable and consistent forecasts.

- Non-regression tests (NRT): Implementation of tests to validate performance stability between different model versions.

- Continuous improvement (Fine-tuning): Refinement of models through fine-tuning to increase their accuracy, robustness, and overall performance.

- Data preparation: Data processing using Python for cleaning, transformation, and normalization before model training.

- Design of automated KPIs: Development of automated KPIs using Tableau and SQL to ensure constant monitoring of model performance.

- Visualization of results: Design of model performance visualizations (predictions, WMAPE), allowing for detailed analysis to inform strategic decisions.

- Automation with Python: Creation of Python scripts to automate recurring processes and optimize efficiency.

- Docker environment management: Configuration and management of Docker environments to ensure model testing under consistent conditions.

- Linux administration: Linux server administration via MobaXterm and task automation with Bash/Shell scripts for smooth resource and system management.

- Data cleaning and preparation: Data Cleansing and Data Preparation with Python to ensure data quality.

- KPI Design: Creation of automated KPIs and interactive, automated dashboards with Tableau Software for performance monitoring.

- Forecast analysis: Evaluation of model performance with metrics such as Accuracy, MAPE, and WMAPE.

- Sales analysis: Study of the impact of holidays and vacations on sales forecasts.

- Automation of calculations: Development of Python scripts to automate the calculation of growth rates and sales distributions.

- Communication in English: Presentation of results in English to stakeholders.