Amine A.

Co-trainer in active pedagogy for a 6-month bootcamp, aiming to qualify a cohort of 14 people at RNCP Level 6 (equivalent to Bachelor's degree). This experience allowed me to contribute to training talent in an inclusive and supportive environment, promoting access to digital skills.

- Python Programming (advanced level)

- APIs and service integration

- SQL Databases (MySQL, PostgreSQL, etc.)

- NoSQL Databases (MongoDB, Cassandra, etc.)

- Data manipulation and DataViz (Numpy, Pandas, Matplotlib, Dash)

- Machine Learning and Deep Learning

- Big Data (Spark, Kafka, Cloud/Azure)

- DevOps (Docker, CI/CD, Git, GitHub Actions)

- MLOps (MLFlow, AirFlow, etc.)

Assignment within a software publisher dedicated to independent healthcare professionals, recognized as the leader in the production and marketing of management, teletransmission, and continuity of care solutions, with a 30% market share.

- Development and production deployment: I led the development of the first projects, establishing robust data pipelines and applying advanced engineering practices to ensure the reliability and efficiency of the solutions.

- Architecture and technological choices: I designed the data infrastructure by integrating solutions like Docker for containerization, MLFlow for model tracking and versioning, and Amazon S3 for massive data storage, enabling optimal management of critical healthcare data.

- Recruitment and team management: I was responsible for recruiting and managing a team of 2 Data Scientists and 1 MLOps engineer, thereby strengthening the company's capabilities in developing and industrializing data solutions.

- Data extraction via OCR: Implementation of advanced Natural Language Processing (NLP) and OCR techniques to structure medical information from digitized documents, facilitating the work of healthcare professionals.

- Reimbursement detection: Development of data pipelines to analyze transactions related to Social Security and mutual insurance reimbursements, including claim management, to improve practitioners' cash flow.

Within BNP Paribas' Market Making division, I was responsible for the entire data pipeline to develop and deploy financial product recommendation models. My role involved designing and optimizing data pipelines using Spark and Dask to process massive volumes of time series data, as well as performing advanced feature engineering to improve data quality and model accuracy.

I also implemented model versioning processes and rigorous testing to ensure the reliability and reproducibility of results. The models were then integrated into CI/CD workflows, enabling automated and iterative production deployments.

The developed solutions were used by:

- Salespeople to optimize inventory and sales campaigns, thanks to accurate customer scores and sales forecasts based on predictive analysis.

- Traders to anticipate market demand risks, by providing real-time recommendations to adjust positions.

Keywords: Spark, Dask, data pipelines, feature engineering, model versioning, model testing, CI/CD, time series, customer scoring, sales forecasting, financial product recommendation.