Oumar KEITA

Topic: Characterization of phosphate groups relative to the K09 reference and prediction of their impact on chemical performance. During my graduation internship at OCP Group, I had the opportunity to apply data science and machine learning techniques to optimize industrial chemical processes. - I developed a conformity index — the K09 Vector Conformity Index (KVCI) — for assessing phosphate quality relative to the K09 reference. The KVCI was built using orthogonal projection, cosine similarity, and Euclidean distance. - I applied unsupervised learning algorithms to segment phosphate samples into four clusters, identifying the group with the highest KVCI as the most suitable for chemical processing, resulting in a potential 4.9% reduction in sulfuric acid consumption. - I designed a predictive impact model that achieved an R² score of 0.88, compared to 0.10 for the existing model, by leveraging advanced machine learning techniques and chemical feature engineering. - Finally, I deployed the model using FastAPI and Docker, enabling reproducible and scalable inference for real-time evaluation of phosphate profiles.

During my research internship in Mexico, I had the privilege to work with Bodesa, a mexican retail company, in collaboration with Colima's State University. As part of the internship : - I analyzed data from 55 departments of Bodesa to identify trends for the remaining 6 months of 2024 and deduced the statistical properties of sales times series for each department - I built forecasting models for each of the 55 departments to predict monthly customer demand leveraging times series forecasting models and AutoML (Automated Machine Learning). - I worked with cross-functional teams to maintain high customer satisfaction and prevents stockouts by accurately forecasting demand.

As communication manager, I am responsible for implementing effective communication strategies to promote the activities of the E-MUN club.

I play a central role in sharing and developing skills in data processing and analysis within the club.

I shadowed an assistant professor in the MSN department for a week. I was exposed to the real working conditions of a professional. I also observed how a leader interacts with his subordinates and all those around him

During my two-month internship, we drew up a monitoring and predictive maintenance plan for BLUEDOVE's camera and other systems, and also started developing the new version of BLUEDOVE's No Waiting UM6P application.