Associate Data Scientist
Societe Generale · Jun 2022 – Apr 2026
- Designed and deployed a production-grade 3 step binary Logistic Regression model for financial crime risk classification, supporting KYC and regulatory compliance.
- Built scalable ML pipelines using Python, PySpark, SQL, and Hive, integrating structured and unstructured data with feature engineering and data preprocessing.
- Developed and evaluated Logistic Regression, Random Forest, XGBoost, K-Means Clustering, and anomaly detection models using Scikit-learn, selecting models based on AUC-ROC and F1-Score.
- Reduced false-positive spoofing alerts by 77% through a hybrid K-Means + NLP solution, improving analyst productivity.
- Applied ANN, CNN, TensorFlow, and Keras for deep learning projects involving classification and model experimentation.
- Conducted 15+ production model monitoring cycles using PSI, model drift, and performance metrics to ensure model stability and regulatory compliance.
- Collaborated with cross-functional teams to support model validation, deployment, governance, and stakeholder reporting.

