Zakaria BOULKHIR

Mainly intervening in projects including data migration, automation of banking processes and data governance. I had learned very valuable financial concepts, data governance techniques and also business analysis (functional business development).

In a large team of data scientists, data analysts, data engineers and MLOps, I was tasked with developing a recommender system at Leroy Merlin’s premises: - Write and publish (internally) a research report on the state-of-the-art. - Develop a Matrix factorisation model that takes collaborative filtering techniques to account. - Deploy the model on the cloud in collaboration with MLOps.

In a microstructure laboratory in MIT engineering school, I worked with a professor who develops generative IA algorithms. My tasks were: - Developing a GAN model from state-of-the-art. - Enhancing the performances using elaborate techniques of the training process and data manipulation (Progressive growing GANs). - Report on the performance enhancement and the choices of models in a research paper.

First ML hands-on project with a research professor in Inria. Tasked with: - Studying state-of-the-art clustering algorithms. - Developing DBSCAN model (probability based) in Typescript. - Adjusting datasets (points of interests vs segments). - Adding the principal of K-anonymity.

In a team of 4 researchers, I was tasked with: - Maintaining the structure of the website amiunique.org. - Adding new significant parameters to fingerprints. - Performing a study with the +1 millions data points that the website have recorded (for classification purposes).