Project Global freight management solutions and services, specializing in Freight Audit & Payment, Order Management, Supplier Management, Visibility, TMS and Freight Spend Analytics.
Overview We are looking for a Data Scientist with strong background in statistics and probability theory to help us build intelligent analytical solutions. The current focus is on outlier detection in freight management data, with further development toward anomaly detection and forecasting models for logistics and freight spend. The role requires both deep analytical thinking and practical hands-on work with data, from SQL extraction to model deployment.
Key Responsibilities * Apply statistical methods and machine learning techniques for outlier and anomaly detection. * Design and develop forecasting models to predict freight costs, shipment volumes, and logistics trends. * Extract, preprocess, and transform large datasets directly from SQL databases. * Categorize exceptions into business-defined groups (e.g., High Value Exceptions, Accessorial Charge Exceptions, Unexpected Origin/Destination). * Collaborate with business analysts to align analytical approaches with domain requirements. * Use dashboards (e.g., nSight) for validation, visualization, and reporting of results. * Ensure models are interpretable, scalable, and deliver actionable insights.
Requirements * Strong foundation in statistics and probability theory. * Proficiency in Python with libraries such as pandas, numpy, matplotlib, scikit-learn. * Proven experience with outlier/anomaly detection techniques. * Hands-on experience in forecasting models (time-series, regression, or advanced ML methods). * Strong SQL skills for working with large datasets. * Ability to communicate findings effectively to both technical and non-technical stakeholders.
Nice to Have * Experience with ML frameworks (TensorFlow, PyTorch). * Familiarity with MLOps practices and model deployment. * Exposure to logistics, supply chain, or financial data. * Knowledge of cloud platforms (AWS, GCP, Azure).
