ichaoui chaima

Addressed the cold-start problem in fashion retail by estimating demand with a 7% error using similarity modeling and lifecycle classification. Architected a three-agent system with LangGraph and scaled the solution to 1M+ transactions using PySpark and Databricks.

Investigated emotion inference from typing behavior using keystroke dynamics, achieving 85% accuracy with a dual-attention Transformer. Developed a web interface with FastAPI for data collection and real-time emotion prediction.

Developed a Deep Neural Network to estimate individual machine energy consumption from total consumption, achieving R² = 0.89. Prepared time-series inputs by synchronizing operational and energy data and engineering features.

Modernized analytical infrastructure by migrating an OLAP cube from Hyperion Planning to Microsoft SSAS. Developed SSIS ETL pipelines to extract, transform, validate, and load cube data while preserving analytical structure.

Developed a stock forecasting solution using Facebook Prophet, achieving R² = 0.90. Deployed the prediction pipeline through a Flask web application for end users.

Pursuing a degree in Data Engineering and Decision-Making Systems, expected to complete in 2026.

Completed a Bachelor's degree in Data Analysis and Big Data.

Earned a Baccalaureate in Experimental Science.